Description
Implementing Self-Service BI to Improve Business Decision Making
Implementing Self-Service BI to
Improve Business Decision Making
Learn how the Intel Technology Manufacturing Engineering organization is using Microsoft BI tools
to help drive innovation at the speed of Moore’s Law.
Executive Summary
Self-service business intelligence (BI) represents a paradigm shift in the way businesses use their data. By
enabling end users to combine and analyze large amounts of data, self-service BI provides a pathway to better
decision making across the enterprise. Yet effective decision making for core business processes often requires
trusted reports that are best developed by centralized teams of experts who can validate data quality and
deliver more sophisticated analyses and reports.
This white paper explores the approach used by the Intel Technical Manufacturing Engineering (Intel TME)
organization to balance these very different requirements using a Microsoft BI solution stack based on SQL
Server*, SharePoint*, and Excel*. By integrating self-service BI with a centralized BI development model, the
Intel TME BI team is creating a more fexible and effcient environment for supporting end-to-end information
needs.
The Microsoft solution stack simplifes this integration by enabling consistent tabular data models to be used
across the entire BI environment. Development teams can deliver plan of record (POR) reports and dashboards
that users can rely on for accuracy and reliability. Individual users and small teams can use those data models
and reports or create their own. They can also adapt and extend them as needed within their personal
“sandbox” environments, including adding data from other sources and altering existing data models.
The Intel TME BI team is evolving its environment to integrate these new capabilities. The result is an
increasingly agile approach to information delivery that is helping employees get the information they need to
manage one of the world’s most complex capital supply chains more effectively. This white paper describes the
solutions and processes the Intel TME BI team is using to implement self-service BI to the hundreds of users
within the Intel TME organization. The paper also discusses how Intel IT is using a similar approach to support
thousands of additional users in its broader business environment.
Authors
Kalpesh Shah, Software Engineer, Technology Manufacturing and Engineering
Eduardo Gamez, Software Engineer, Technology Manufacturing and Engineering
David Yantis, PMP, SPBI Product Manager, Information Technology (IT)
Rob Shiveley, Software Alliance Marketing Manager, Software and Solutions Group
WHITE PAPER
Intel® Xeon® processor E5 family
Business Intelligence
Table of Contents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Coordinating a Global Technology Ecosystem . . . . . . . . . . . . . .1
The Information Challenge: Scattered Data
and Static Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Implementing Next-Generation BI . . . . . . . . . . . . . . . . . . . . . . . .3
New Technologies Create New Opportunities . . . . . . . . . . . . . . . .3
Implementing the Microsoft BI Solution Stack . . . . . . . . . . . . .4
Ensuring Data Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Unifying BI across All User Groups . . . . . . . . . . . . . . . . . . . . . . . . . . .6
Optimizing BI for Agility and Effectiveness . . . . . . . . . . . . . . . .6
Targeting Big Needs with a Centralized Team . . . . . . . . . . . . . . . .6
Extending Value with BI “Soft-Serve” . . . . . . . . . . . . . . . . . . . . . . . .8
The Need for Speed – Improving Velocity through
Self-Service BI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
A Decentralized Approach to Training and Support . . . . . . . 10
Moving Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
Success to Date . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Future Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Additional Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Coordinating a Global Technology Ecosystem
In 1965, Intel founder Gordon Moore published a paper predicting
that the number of transistors that could be manufactured on
a chip would continue to double every two years. That pace of
innovation now provides the foundation for Intel’s Tick-Tock
model of development, in which Intel alternates new silicon
manufacturing technologies with new microarchitecture designs
to deliver new products on an annual timetable.
This fast, predictable cadence is key to Intel’s success and to
the success of many Intel customers. Yet every new product
generation introduces greater design complexity and a slew of
new technology demands. Lithography must be more precise and
new materials and manufacturing processes are often needed,
all of which require new fabrication equipment and processes.
Multiply these demands across Intel’s global manufacturing
footprint, which currently includes 11 major manufacturing
sites, and the task of continually upgrading and managing Intel’s
capital assets becomes a very complex undertaking.
The Intel Technical Manufacturing Engineering (TME)
organization is responsible for keeping this process on track,
and for doing it affordably. This group of approximately 600
employees manages Intel’s capital supply chain across fve
technology generations (Figure 1). Intel TME’s responsibility for
each generation begins with research and development and
continues on through design, manufacture, installation, use,
maintenance, and end-of-life management.
Intel TME works with hundreds of internal and external groups,
including commercial and academic research teams, tool design
and manufacturing vendors, and many others. The decision
matrix for keeping core processes on track is complex. Issues
such as technology, velocity, sustainability, and affordability
must be balanced in every decision.
COLLABORATIVE
RESEARCH
INTERNAL
RESEARCH
DEVELOPMENT MANUFACTURING END OF LIFE
Leveraging
Investments
Securing the Right
Technologies
Choosing the Right
Partners
Delivering to the Core
Extending the Value
Spending Money
Wisely
Extracting Full
Value
Figure 1. From research and development to end-of-life, the Intel Technology Manufacturing and Engineering group manages the lifecycle of Intel’s
capital manufacturing equipment to deliver affordable innovation at the speed of Moore’s Law.
2
Implementing Self-Service BI to Improve Business Decision Making
The Information Challenge: Scattered Data and Static Reports
Intel TME generates large amounts of data in managing Intel’s capital supply chain. Much of this data is scattered throughout the
organization. It resides within production applications, databases, and data warehouses at the departmental and enterprise levels,
and also within Excel spreadsheets and Microsoft Access* databases owned and managed by individual employees. In most cases,
employees use Excel and manual methods to process, update, and maintain the data, and they create presentations on an ad hoc basis
by copying and pasting Excel tables and graphs into Microsoft PowerPoint* presentations (Figure 2).
Until recently, there has been no unifed focus on standardizing
supply chain data and validating its quality across the Intel
TME organization. However, the pace of business continues to
accelerate at Intel, as it does throughout the business world.
Faster access to data and better, faster decision making is
increasingly necessary to keep pace with demands, especially
as complexity continues to climb. It is no longer tenable to
take months to deliver data and information. Better tools and
methods are needed.
Implementing Next-Generation BI
In most cases, the frst step in improving decision making is to
speed “time-to-data.” This enables individuals to devote more
of their time and effort to exploring and analyzing the available
data before their decision window closes (Figure 3). Ideally, this
results in both faster and better decisions that deliver higher
value to the business.
Other Data
Source(s)
Microsoft
SQL Server*
Individual Access
Business User, Data Analyst,
Commodity Manager, etc.
Microsoft
Excel*
Microsoft
PowerPoint*
Figure 3. The first step in improving decision making is speeding
“time-to-data,” so employees can focus more of their time and effort
on analyzing the available information to make higher quality decisions
in less total time.
Figure 2. In most cases, Intel Technology Manufacturing and Engineering employees currently use Microsoft Excel* and manual methods to process,
update, and maintain data, and they create presentations by copying and pasting Excel tables and graphs into Microsoft PowerPoint* presentations.
TIME-TO-DATA
TIME-TO-DATA
TIME-TO-DECISION
TIME-TO-DECISION
TIME-TO-
INFORMATION
TIME-TO-
INFORMATION
DECISION
DECISION
New Technologies Create New Opportunities
A number of technologies and solutions have emerged recently
that can help organizations speed time-to-data for individual
decision makers. Two new capabilities stand out as particularly
relevant.
• Real-time analytics acting on large data sets. This capability
is typically delivered by a combination of several database
technologies, such as in-memory analytics, columnar data
structures, and data compression. These technologies take
advantage of the parallel execution capabilities and large
memory capacities delivered in recent server generations.
For example, a four-socket server based on the Intel® Xeon®
processor E7 family
1
now provides up to 80 execution threads
and supports up to 2 terabytes of memory. With these
resources, even very large data sets can be held entirely in
memory and large numbers of concurrent queries can be
executed in parallel.
• Self-service BI. Historically, an end user interested in accessing
business data and creating visualizations would require end-to-
end assistance from IT staff. A new class of easy-to-use BI tools
is now available that allows typical business decision makers to
quickly access, combine, and analyze data, create presentations,
and then share the results with others.
In combination, these innovations allow IT organizations to shift
their focus. Instead of delivering reports, they have the option
of delivering data models and visualization tools to the business
and allowing users to create their own reports. Although there
are many sophisticated analytic techniques that still require
expert data analysts and developers, these self-service tools
fll a critical gap to enable more effcient and fexible delivery
of enterprise information. 3
Implementing Self-Service BI to Improve Business Decision Making
3
The Intel TME implementation of the Microsoft BI solution
stack is shown in Figure 4. The back end of the BI environment
consists of an organizational data mart built on SQL Server. The
team uses a number of built-in services to manage the data
mart, including the following:
• SQL Server Integration Services (SSIS) for deploying extract,
transform, and load (ETL) packages to configure data feeds
from external sources
• SQL Server Analysis Services (SSAS) for deploying and
managing cubes and tabular models within the data mart
• SQL Server Reporting Services (SSRS) for creating reports
• Microsoft Visual Studio* with BI Development Studio (BIDS)
and SQL Server Data Tools (SSDT) for development teams
to write BI applications, including ETL packages, data models,
reports, and dashboards
Implementing the Microsoft BI Solution Stack
Intel TME’s initial drive to improve BI came from two directions.
The lengthy timelines associated with traditional BI projects
were delaying business access to information and also placing a
heavy burden on BI development teams. A business analyst in
the Intel TME capital planning organization was engaging with
key staff to address this issue. At the same time, a software
developer in the Intel TME Systems Engineering group was
exploring technical solutions using the Microsoft BI stack and
Kimball data warehousing methodologies.
The employees driving these independent efforts connected,
providing a fortuitous combination of need and solution. Initially
the two employees focused on a single business need: measuring
supplier delivery performance. This frst project delivered
suffcient business value to justify the creation of a formal Intel
TME BI team to develop and execute a roadmap.
The Microsoft BI solution stack aligns well with Intel needs
and requirements. Most Intel TME data resides in SQL Server
databases and multi-dimensional cubes, personal Microsoft
Access databases, and Excel spreadsheets. Microsoft BI tools
make it easier to take advantage of these data sources and to
leverage existing knowledge and skillsets.
INTEL TME BI ARCHITECTURE
Tabular Mode
Multidimensional Mode
Microsoft SharePoint
Hardware
Infrastructure
Microsoft SQL Server
Analysis Services (SSAS)
Excel Services
Power View for SharePoint
(Reporting Services)
PowerPivot for SharePoint
(Analysis Services)
Microsoft
Internet Explorer
PowerPivot
for Excel
Microsoft Excel
Third Party Apps
Team BI Personal BI Organizational BI
Storage
Clustered
Virtual Hosts
Virtual
Machines
4 vCPUs, 12GB memory
(production, test, development)
16TB shared SAN
5 ProLiant BL490c G7 Blades
(2 x Intel® Xeon®processor 5600 series and 400GB memory per blade)
4vCPUs, 6GB memory
(production, test, development)
Figure 4. The Intel Technology Manufacturing and Engineering BI implementation provides users with flexible access to data, reports, and BI tools
using a Microsoft BI solution stack that includes Excel*, Internet Explorer*, SharePoint*, and an organizational data mart built on Microsoft SQL Server*.
4
Implementing Self-Service BI to Improve Business Decision Making
UPGRADING THE SOFTWARE STACK TO DELIVER HIGHER VALUE
The Intel Technology Manufacturing and Engineering BI solution stack currently includes Microsoft SQL Server* 2012,
SharePoint* 2010, Windows Server* 2008 R2, and Excel* 2010, as well as System Center 2007 for assisting infrastructure
management teams. Intel is currently in the testing and planning stages of an upgrade that will include moving to SharePoint
2013, Windows Server 2012, Excel 2013, and System Center 2012. The upgrade will most likely include a move to new blade
servers based on the latest Intel® Xeon® processor E5 family to reduce the data center space and power footprint.
There are many advantages to this updated solution stack. One of the most valuable for Intel TME is that PowerPivot and
Power View will be fully integrated across all software components, so tabular data models will be supported throughout the
BI environment, with no need for plug-ins or workarounds. This will allow the team to provide users with direct access to data
models within the Intel TME data mart to support full self-service BI functionality. As part of the move to Windows Server 2012,
the team is currently testing the latest version of Hyper-V*, and anticipates being able to reduce software licensing costs for
supporting the virtualized server environment.
For more information about Microsoft BI solutions and the Intel Xeon processor E5 family,
see the resource links at the end of this paper.
SharePoint provides the foundation for managing end-user
access to data and libraries of shared reports. End users have
the option of logging onto SharePoint with their browsers and
using server-side instances of Excel to view and interact with
data and reports. Alternatively, they can download data and
reports for local execution on their personal instances of Excel.
(Note: The Intel TME BI team currently provides access to shared
reports. Providing users with direct access to data models is a
work in progress. See the sidebar, Upgrading the Software Stack
to Deliver Higher Value.)
The Intel TME BI team manages access rights using Microsoft
Active Directory*. The security model extends into SQL Server,
which enables granular control of data access. This is important,
since the Intel TME data mart contains some of Intel’s most
sensitive intellectual property (IP).
The hardware infrastructure for the Intel TME BI implementation
consists of fve clustered, two-socket blade servers, each
confgured with two six-core Intel® Xeon® processors and 400 GB
of memory. These servers currently host nine virtual machines:
two for the production environment, two for staging, and fve
for development. All instances of SQL Server run in a single
virtual machine confgured with 4 vCPU and 6 gigabytes (GB)
of memory. All instances of SharePoint Server run in a single
virtual machine confgured with 4 vCPU and 12 GB of memory.
Infrastructure teams are currently exploring the advantages of
increasing memory to 16 GB for the SharePoint virtual machines.
The large amount of system memory is provided to support
Microsoft xVelocity In-Memory Technology (discussed below),
which provides fast data manipulation and queries by holding
all data in memory, rather than shuttling data back and forth
between memory and disk.
Ensuring Data Quality
Data quality and completeness are critical success factors for
all BI efforts. Most data in the Intel TME data mart comes from
Intel TME production applications and databases. As the Intel
TME BI team starts processing data for BI consumption, they
also perform basic auditing to ensure validity. This is especially
critical for data that will be used for measures and indicators.
Since data is consumed differently for BI than it is in other use
cases, auditing often uncovers data quality issues that would
most likely go unnoticed without these efforts. Once an issue is
identifed and a root cause is determined, a project is launched to
close the data quality gap.
5
Implementing Self-Service BI to Improve Business Decision Making
5
The Intel TME BI team uses several tools to audit data, including
those provided in SQL Server 2012. The team recently began
providing select business users with direct access to the data
mart’s staging environment, facts, and dimension tables using
Excel. They have found that business users provide valuable
insights into data quality and usability.
Unifying BI across All User Groups
The Intel TME data mart was initially deployed on SQL Server
2008 R2. However, a recent upgrade to SQL Server 2012 has
delivered important new features. Combined with capabilities
available in Excel and SharePoint, these enhancements are
key to delivering effective self-service BI and integrating it
effciently with centralized BI development efforts.
• The Microsoft BI Semantic Model (BISM) in SQL Server 2012
includes new tabular data models in SSAS, as well as more
traditional multidimensional models. These tabular models can
be accessed and consumed directly by business decision makers
using PowerPivot and Power View.
• PowerPivot allows end users to manipulate large pivot tables
and tabular data models at high speed within an Excel workbook.
High performance and scalability are enabled by Microsoft
xVelocity In-Memory Technology, which combines in-memory
analytics processing with columnar data models and advanced
data compression. PowerPivot is available as a plug-in for Excel
2010, and is supported in SharePoint 2010. Microsoft has also
integrated xVelocity In-Memory Technology into SQL Server
2012 to accelerate performance for server-side data analytics.
• Power View is a self-service reporting tool for business users
that is integrated into SharePoint 2010 and into SQL Server
2012 (as part of SSRS). Power View provides an interactive
interface for users to explore, manipulate, and present the
data located in tabular data models and pivot tables. Both
PowerPivot and Power View will be included in Excel 2013,
which the Intel TME BI team is currently testing and plans to
deploy in conjunction with SharePoint 2013.
With BISM, PowerPivot, and Power View, tabular data models
are now supported across the entire Microsoft BI solution stack,
including SQL Server, SharePoint, and Excel. This consistency
enables a unifed BI environment across all usage models.
For Intel TME, this is one of the primary advantages of the
Microsoft BI solution. It is now relatively simple and seamless
for development teams to:
• Create BISM tabular models using Visual Studio SQL Server
Data Tools (SSDT). These models are then deployed to SSAS
tabular instances to be consumed by Power View reports.
• Use the Data Analysis Expressions (DAX) language within
tabular models to quickly create and deploy both simple
and extremely complex data aggregations. Unlike SSIS, DAX
provides a very efficient way to create new measures and
indicators and quickly deploy them from development, to
testing, to production environments.
With this foundation, centralized teams have an effcient BI
development framework for developing tabular data models.
End users can potentially access, manipulate, and consume those
same data models using Excel with PowerPivot and Power View.
The result is a fexible and unifed BI environment capable of
supporting the full range of Intel TME BI requirements.
Optimizing BI for Agility and Effectiveness
Like most organizations, Intel TME has a continuum of data
needs. At one end are the “big deals” that are directly tied
to core business processes and objectives, such as vendor
assessment and procurement tracking. These processes depend
on large amounts of scattered data and have a direct and
signifcant impact on Intel’s overall success. At the other end
of the continuum are the data needs of individuals and small
groups. Improving data access for these users would not have a
major impact on core business functions but might, collectively,
have a substantial impact on overall business success.
Targeting Big Needs with a Centralized Team
The Intel TME BI Working Group was formed to drive
organizational alignment on the big deals by initiating and
managing BI deliverables and working to enable successful
integration into Intel TME operations. The team gathers
requirements from each participating department and
develops a prioritized roadmap for implementation by the
Intel TME BI team.
Seven major user groups within Intel TME are represented
within the Intel TME BI Working Group, with responsibilities
ranging from equipment development and capital planning,
to training and factory ramping, to business operations and
capital equipment resale. Most of the representatives are
“data stewards” for their organizations. These data stewards
are typically Excel power users and have a good, practical
understanding of business processes, data sources, and
information needs. A few are experts in data analytics, but
most are not. A few organizations lack anyone with the required
skillsets, in which case advocates are enlisted from other groups.
6
Implementing Self-Service BI to Improve Business Decision Making
Once high-level requirements have been gathered, projects are evaluated and prioritized by the Intel TME BI Working Group. In an
attempt to avoid the kinds of political issues that can arise in cross-organizational decision making, the group developed a standardized
process for evaluating projects based on estimated risk and reward (Figure 5). Projects with low risks and high rewards get top priority
(see the sidebar, Characteristics of a Successful BI Project).
Risk
B
u
s
i
n
e
s
s
V
a
l
u
e
Project A
Project B
Project C
Project D
Project E
Project F
Project G
Project H
Project I
Figure 5. To allocate resources in the most effective manner, the Intel
Technology Manufacturing and Engineering BI Working Group developed
a standardized process for evaluating projects based on estimated risk
and reward.
Depending on the requirements, some combination of database
administrators (DBAs), data analysts, business analysts, and
BI software developers work together on each project. These
teams have found that BI development is very different from
traditional application development. Instead of building a
functional map based on business process needs, developers
must begin with a clear understanding of data sources and then
map business needs to create measures and indicators of the
available data. The process for a typical BI project includes four
key steps.
1. Engage with business teams to develop user stories that
defne requirements, including data, measures, indicators,
and reports. This step typically consumes approximately
three-fourths of the total time and effort for each project,
and the refnement of these requirements typically extends
throughout the development process.
2. Develop ETL packages, tabular data models, and data mart
changes using Microsoft Visual Studio, SSIS, and SSAS. At
this stage of Intel TME’s BI evolution, this step often involves
accessing, auditing, and integrating new data sources.
3. Create the associated reports and dashboards using Power
View. Work with data stewards and their business teams to
test and refne the solution.
4. Release the fnalized reports to production and publish
them on SharePoint as part of the Intel POR.
CHARACTERISTICS OF A SUCCESSFUL BI PROJECT
The Intel Technology Manufacturing and Engineering BI development team has found that active support from targeted
business groups is crucial to the success of BI development projects. When assessing the viability of BI requests from
business units, proposed projects should be:
• Tied to a documented business process with a clearly envisioned end-state.
• Supported actively by business representatives during requirements defnition, data validation and business process
alignment. These phases typically account for about three fourths of the work in BI development, and it is important the
business representatives are available and responsive.
• Championed by one or more key business users to fuel post-rollout adoption and to provide some level
of training and support for novice business users.
7
Implementing Self-Service BI to Improve Business Decision Making
Extending Value with BI “Soft-Serve”
Intel TME POR reports can be published only by the Intel TME BI
team, which follows standardized procedures to maintain quality
and reliability. Decision makers can use Power View to access
these POR reports (Figure 6). They can also copy the reports into
their personal sandbox areas and adapt them in any way they
choose. These altered reports are no longer POR, but they can
be widely shared. Access is limited only by the Active Directory
groups the Intel TME BI team maintains to control access to sites,
sub-sites, and libraries. Since these sandbox reports point to the
same tabular database, the data is secured no matter who views
a specifc sandbox report.
TME Data Mart
Other Data Sources
TME Depts
BI Working Group
Prioritized
Roadmap
Analyst
Analyst
User Stories
Developer
Business User
Additional
BI Requests
Self-Service
BI Opportunity
Business User
Microsoft SharePoint
Microsoft
SQL Server
Reports
“SOFT-SERVE” BI
Figure 6. As an interim step before implementing full self-service business intelligence (BI), the Intel Technology Manufacturing and Engineering
“soft-serve” BI approach allows users to access reports and adapt them to meet their specific requirements.
Intel TME calls this BI model “soft-serve.” Unlike a full self-service
BI model, the soft-serve model does not allow end users to
access and manipulate data from the data mart, only reports.
However, even this interim soft-serve model has substantial
value, since it enables the Intel TME BI team to standardize
data, measures, indicators, reports, and visualizations, while
empowering end users to modify reports as needed to address
their unique requirements.
8
Implementing Self-Service BI to Improve Business Decision Making
The Need for Speed – Improving Velocity through
Self-Service BI
The soft-serve approach to BI helps to improve fexibility for
end users, but not enough to meet the rapidly growing business
demand within the Intel TME organization. Business decision
makers want even greater fexibility, and they don’t want to
wait on lengthy centralized projects to get the information they
want in the way they want it. At the same time, the TME BI team
needs to be able to standardize and centralize TME data and
information to ensure data quality. These competing needs place
unsustainable demands on centralized IT resources. Like IT shops
everywhere, the Intel TME BI team needs a way to standardize
and validate organizational data, while simultaneously delivering
more agile and differentiated support to end users.
To address these needs, Intel TME is working to deliver full self-
service BI to data stewards and other end users. This approach
allows users to access data models within the TME data mart
and create their own reports using Power View. They can also
download the data models into their personal sandbox areas,
after which they have complete fexibility to add or remove
data, restructure models, and create their own relationships,
calculated columns, measures, and key performance
indicators (KPIs).
2
Users can then review and refne their data models and reports
in collaboration with their own business units and make the
reports available to authorized users through SharePoint. When
a user-generated report is mature, shown to deliver high value,
and approved by the appropriate business unit, the Intel TME
BI Working Group can coordinate a Technical Review Group
(TRG) to create a more complete user story, perform data
audits, determine data gaps, and enhance the report to ensure
standardized measures, metrics, and KPIs.
The new data model and report can then be deployed within
the Intel TME data mart and become part of the POR. The cycle
can then repeat, with users continuing to extend and adapt POR
reports within their private sandboxes to address their unique
needs and then sharing the results with others.
This self-service approach to BI represents a paradigm shift in
data and information delivery. It is already impacting the way
Intel TME develops and delivers data and reports. Traditional
agile processes for software development are still valuable for
centralized solutions, but they are being adapted to incorporate
elements of this more decentralized approach. Business unit data
stewards are being integrated into the development process,
and soft-serve and self-service BI is becoming an essential and
integrated part of the overall offering.
Soft-serve and self-service BI introduce risks as well as rewards.
Without centralized control, not all data models and reports will
be high quality. Some may even provide wrong or misleading
information. However, it is Intel’s belief that the potential value
far exceeds the risk by providing an exceptionally broad base
for innovation. Hundreds of users, all with different needs,
experience, and insight into business processes, will be able to
access and analyze data faster and more effectively.
Intel TME is not the only organization within Intel that can
beneft from faster access to data and information, so Intel
is extending access to self-service BI tools throughout the
enterprise. This broader self-service BI implementation uses the
same software stack, but a different physical implementation.
3
It went live in 2011 and had over 5,000 users within one year.
By mid-March 2013, there were 13,800 active users
and the user base continues to grow.
9
Implementing Self-Service BI to Improve Business Decision Making
A Decentralized Approach to Training and Support
Effective use of self-service BI requires that users understand
the following:
• Relevant business processes
• Pertinent data sets, including how they are organized and
interrelated and how to access them
• Tools and methods for analyzing, presenting, and sharing data
It would be possible to train large user groups to use BI tools, but
business processes and relevant data vary across the company.
Because of this, training end users for BI will never be like
training end users to use a traditional business application. A
decentralized approach to BI training helps to address this issue.
PowerPoint
Self-Service BI
TRG
TME BI Roadmap
User Stories
Project Status
Ready for
TME BI?
KANBAN
TME BI Application
LIfecycle Management
Create/Update
Update
Deploy
Review
TME Dept
Microsoft
SharePoint
Yes
No
Data Steward
Data Steward
Data
Steward
Data Steward
Analyst
Developer
Business User
Data Steward
Other
Data Sources
TME
Data Mart
SELF-SERVICE AND CENTRALIZED BI INTEGRATION
Microsoft
SQL Server
Figure 7. Intel Technology Manufacturing and Engineering is moving toward full self-service business intelligence, in which data stewards and other
users can access data as well as reports and adapt both in any way they choose within their personal sandbox environments.
The Intel TME BI team is educating data stewards to use the BI
tools. Intel corporate IT takes a similar approach, training site
owners and select users throughout the business. In both cases,
those individuals then train and support people within their own
organizations. This process helps to get the right information and
training to the right people, by creating expanding “communities
of excellence” tailored to the needs of each organization.
Intel IT offers some additional training sessions on an ad hoc
basis to help users address common challenges. Users can also
attend online “offce hours” to ask specifc questions that may
be beyond the expertise of their data stewards. For now, this
strategy is working well, enabling a small team to train and
support a very large user group. As self-serve BI continues to
increase in popularity and importance at Intel, it may become
necessary to deliver more formal training to supplement the
current programs.
10
Implementing Self-Service BI to Improve Business Decision Making
Moving Forward
Success to Date
The Intel TME soft-serve and self-service BI programs are still
in a relatively early stage. The Intel TME BI team released one
centralized report to production in 2011 and fve more in 2012.
Many more are in the development queue for 2013. One example
of a successful BI deliverable is the Intel TME Forecast Progress
Indicators report, which is used to help Intel TME stakeholders
understand the current status of the capital supply chain.
Intel TME has never before had a single view showing the status
of all the tools at each supply chain stage, much less with this
level of detail. The report provides a summary page showing
the key steps—Forecast, Order, Delivery, Impact, Transit, and
Install—and provides drill down alternatives to view graphs and
explore details. Gaps can be identifed quickly and explored to
help mitigate risks and keep factory ramping schedules on track.
This information is now delivered automatically to all
stakeholders, so it is no longer necessary to create and run
manual reports. The Intel TME Forecast Progress Indicators
report and several others have reduced time-to-data to help
business decision makers monitor, evaluate, and manage key
processes, including the following:
• Supplier delivery performance
• Procurement status across the capital supply chain
• Risk factors for equipment implementation and quality
• Document readiness
In one case, the delivered report provided approximately 400
hours per quarter in productivity savings. In general, however,
quantifying value in the BI arena is challenging. Individual
decisions must be tracked to determine if they were, in fact, the
best decisions. The speed and quality of the decision must also
be tied back to the BI environment.
Because of these challenges, Intel TME is currently relying on
user perception and feedback to evaluate success. Feedback has
been positive across the majority of projects:
“ In the past, we had to pull and link all this data together
manually from the different systems. The process was so time-
intensive, we only did it once per quarter—at most. Now it takes
just an hour or two to pull and manipulate the data.”
“ This tool has facilitated an in-depth analysis at the SRC line
item level, allowing us to analyze point allocation and actual
scores on the spot. We couldn’t do this before.”
“ For years, I’ve been asking how we could measure work
profciency. Now I have an answer. Great job! ”
“ Thank you for your support in generating tailored data
acquisition pulls and metrics for measuring continuous
improvement. We’ll be happy to use these indicators!!
And more!!! ”
The majority of Intel business users also report an excellent
experience using Excel, PowerPivot, and Power View for
accessing and creating reports and dashboards. Even very large
data sets can be combined, fltered, and processed with fast
application response times and building reports is relatively
simple and intuitive. The ability to continue manipulating data
and visualizations within existing presentations has proved
valuable in creating impactful and easily customizable reports.
This feedback is valuable, but more quantitative data is needed
for effective planning. Efforts are underway to introduce these
and other success metrics in 2013. The primary focus will be on
quantifying the impact reports have on time-to-data. Once the
team has a handle on these metrics, they will begin focusing on
more complex issues, such as decision quality and the ability to
make more decisions in less time.
Future Plans
Although the BI program has delivered clear benefts to date, the
Intel TME BI team believes the biggest beneft is the groundwork
that has been laid for future BI innovation by standardizing
tools and processes for data access, analysis, and visualization.
Processes are becoming more repeatable and success more
predictable. The amount of high-quality data in the Intel TME
data mart is also growing, which will create new opportunities
and faster paths to success going forward.
BI and analytics are a work in progress at Intel, as they are for
every company. The Intel TME vision for BI is a world in which
the following criteria are satisfied:
• Every employee has 24x7x365 access to the data and
information they need to do their job, through a single portal
that is available worldwide and configured to match their
unique role in the business.
• Portal usage is intuitive, with self-serve data and no need
for training.
• Barriers to data access are minimized within the constraints
dictated by the need to protect Intel IP.
• All data accessed by decision makers comes from a single
system of record to ensure data quality and a single version of
the truth. To meet this need, Intel TME must be able to access
all required data across the Intel Technology Manufacturing
Group
4
(TMG) and the Intel TMG supplier ecosystem. This, in
turn, requires that Intel TME data should be fully aligned with
supplier data obtained through the next-generation Intel TMG
Supplier Presence Site.
• BI tools are embraced by all levels of the organization, to
the extent that individuals and teams update their business
processes to make better use of this new information source.
In time, manual sources of information should disappear
completely, because they are no longer needed.
11
Implementing Self-Service BI to Improve Business Decision Making
Additional Resources
For more information about:
• The Microsoft BI solution stack, visit:
www.microsoft.com/en-us/bi/Capabilities.aspx
• Intel® Server Products, including the Intel Xeon processor E7
family, the Intel Xeon processor E5 family, and the Intel® Xeon®
processor E3 family, visit: www.intel.com/content/www/us/
en/processors/xeon/xeon-processor-e7-family.html
It should be emphasized that these are goals, not expectations.
Existing data sets are growing and data sources are proliferating
throughout the business. Fully controlling data and information
will most likely never be a practical possibility. Nevertheless, a
vision of an ideal BI environment helps to guide long-term plans
and has helped the Intel TME BI team establish the following
goals for 2013.
• Manage centralized BI projects as a portfolio. Advances
in quantifying BI project success will help in this endeavor,
providing a better foundation for prioritizing projects and
allocating resources based on expected returns.
• Move toward trending and predictive analytics. Efforts to
date have focused primarily on tracking current activities with
reference to goals and expectations. The team has performed
a few “what if” analyses to try to predict outcomes for better
decision making and sees substantial value in moving toward
this more advanced analytics model. Efforts will continue to
evolve as the TME BI team builds a better understanding of
both technical and business opportunities and hurdles.
• Strengthen BI alignment across Intel. Collaborative efforts
are underway with other groups to define self-service BI at
Intel. As one example, the TME BI team is partnering with the
SharePoint Business Intelligence enterprise team, which is
building an Intel-wide Community of Excellence program around
SharePoint BI.
Conclusion
Evolving a global supply chain at the speed of Moore’s Law is
a complex undertaking, requiring tight coordination among
hundreds of companies and thousands of individuals. Intel TME is
using an integrated combination of centralized and self-service
BI development to provide decision makers with simpler and
faster access to the information they need to do their jobs more
effectively. Microsoft BI tools provide the software foundation
for this integration, with an organizational data mart built on SQL
Server, a managed BI portal built on SharePoint, and end-user
tools provided by Excel (with PowerPivot and Power View)—all
running on Intel® processor-based servers and laptops.
The ability to use consistent data models for both end-users
and centralized development teams has been instrumental in
providing fexible and effective support for all user groups.
Business users can access centrally developed and validated POR
reports to improve decision making for core business functions.
They also have the fexibility to adapt and extend those reports
with new data and alternative analyses, or to build entirely new
data models and reports. As information requirements continue
to grow, the Intel TME BI team believes this strategy will enable
continual improvements in data usage and decision making
throughout the organization.
Implementing Self-Service BI to Improve Business Decision Making
1
For more information on the Intel
®
Xeon
®
processor E7 family, visit the Intel web site. www.intel.com/content/www/us/en/processors/xeon/xeon-processor-e7-family.html
2
Power View reports created in Microsoft Excel* instances currently cannot be imported directly into Microsoft SharePoint* and consumed using reporting tools within SQL Server* Reporting Services, even though the PowerPivot and tabular
data models are typically the same. Technology Manufacturing and Engineering IT is currently exploring options for eliminating this gap.
3
Both the Technology Manufacturing and Engineering implementation and the broader, self-serve implementation use the Microsoft BI solution stack. However, there may be differences in product generations and version numbers.
4
The Intel Technology Manufacturing Engineering (TME) organization exists within the larger Intel Technology Manufacturing Group (TMG) , which is responsible for all aspects of Intel’s global manufacturing assets.
INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL
®
PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE,
TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL’S TERMS AND CONDITIONS OF SALE FOR SUCH
PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO SALE AND/OR USE OF INTEL
PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT,
COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. UNLESS OTHERWISE AGREED IN WRITING BY INTEL, THE INTEL PRODUCTS ARE NOT DESIGNED NOR
INTENDED FOR ANY APPLICATION IN WHICH THE FAILURE OF THE INTEL PRODUCT COULD CREATE A SITUATION WHERE PERSONAL INJURY OR DEATH MAY OCCUR.
Intel may make changes to specifcations and product descriptions at any time, without notice. Designers must not rely on the absence or characteristics of any features or instructions
marked “reserved” or “undefned.” Intel reserves these for future defnition and shall have no responsibility whatsoever for conficts or incompatibilities arising from future changes to
them. The information here is subject to change without notice. Do not fnalize a design with this information.
The products described in this document may contain design defects or errors known as errata which may cause the product to deviate from published specifcations. Current
characterized errata are available on request. Contact your local Intel sales offce or your distributor to obtain the latest specifcations and before placing your product order. Copies
of documents which have an order number and are referenced in this document, or other Intel literature, may be obtained by calling 1-800-548-4725, or by visiting Intel’s Web site
at www.intel.com.
Copyright © 2013 Intel Corporation. All rights reserved. Intel, the Intel logo, and Xeon are trademarks of Intel Corporation in the U.S. and other countries.
*Other names and brands may be claimed as the property of others. Printed in USA 0413/RS/MESH/PDF Please Recycle 328830-002US
doc_965604906.pdf
Implementing Self-Service BI to Improve Business Decision Making
Implementing Self-Service BI to
Improve Business Decision Making
Learn how the Intel Technology Manufacturing Engineering organization is using Microsoft BI tools
to help drive innovation at the speed of Moore’s Law.
Executive Summary
Self-service business intelligence (BI) represents a paradigm shift in the way businesses use their data. By
enabling end users to combine and analyze large amounts of data, self-service BI provides a pathway to better
decision making across the enterprise. Yet effective decision making for core business processes often requires
trusted reports that are best developed by centralized teams of experts who can validate data quality and
deliver more sophisticated analyses and reports.
This white paper explores the approach used by the Intel Technical Manufacturing Engineering (Intel TME)
organization to balance these very different requirements using a Microsoft BI solution stack based on SQL
Server*, SharePoint*, and Excel*. By integrating self-service BI with a centralized BI development model, the
Intel TME BI team is creating a more fexible and effcient environment for supporting end-to-end information
needs.
The Microsoft solution stack simplifes this integration by enabling consistent tabular data models to be used
across the entire BI environment. Development teams can deliver plan of record (POR) reports and dashboards
that users can rely on for accuracy and reliability. Individual users and small teams can use those data models
and reports or create their own. They can also adapt and extend them as needed within their personal
“sandbox” environments, including adding data from other sources and altering existing data models.
The Intel TME BI team is evolving its environment to integrate these new capabilities. The result is an
increasingly agile approach to information delivery that is helping employees get the information they need to
manage one of the world’s most complex capital supply chains more effectively. This white paper describes the
solutions and processes the Intel TME BI team is using to implement self-service BI to the hundreds of users
within the Intel TME organization. The paper also discusses how Intel IT is using a similar approach to support
thousands of additional users in its broader business environment.
Authors
Kalpesh Shah, Software Engineer, Technology Manufacturing and Engineering
Eduardo Gamez, Software Engineer, Technology Manufacturing and Engineering
David Yantis, PMP, SPBI Product Manager, Information Technology (IT)
Rob Shiveley, Software Alliance Marketing Manager, Software and Solutions Group
WHITE PAPER
Intel® Xeon® processor E5 family
Business Intelligence
Table of Contents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Coordinating a Global Technology Ecosystem . . . . . . . . . . . . . .1
The Information Challenge: Scattered Data
and Static Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Implementing Next-Generation BI . . . . . . . . . . . . . . . . . . . . . . . .3
New Technologies Create New Opportunities . . . . . . . . . . . . . . . .3
Implementing the Microsoft BI Solution Stack . . . . . . . . . . . . .4
Ensuring Data Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Unifying BI across All User Groups . . . . . . . . . . . . . . . . . . . . . . . . . . .6
Optimizing BI for Agility and Effectiveness . . . . . . . . . . . . . . . .6
Targeting Big Needs with a Centralized Team . . . . . . . . . . . . . . . .6
Extending Value with BI “Soft-Serve” . . . . . . . . . . . . . . . . . . . . . . . .8
The Need for Speed – Improving Velocity through
Self-Service BI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
A Decentralized Approach to Training and Support . . . . . . . 10
Moving Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
Success to Date . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Future Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Additional Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Coordinating a Global Technology Ecosystem
In 1965, Intel founder Gordon Moore published a paper predicting
that the number of transistors that could be manufactured on
a chip would continue to double every two years. That pace of
innovation now provides the foundation for Intel’s Tick-Tock
model of development, in which Intel alternates new silicon
manufacturing technologies with new microarchitecture designs
to deliver new products on an annual timetable.
This fast, predictable cadence is key to Intel’s success and to
the success of many Intel customers. Yet every new product
generation introduces greater design complexity and a slew of
new technology demands. Lithography must be more precise and
new materials and manufacturing processes are often needed,
all of which require new fabrication equipment and processes.
Multiply these demands across Intel’s global manufacturing
footprint, which currently includes 11 major manufacturing
sites, and the task of continually upgrading and managing Intel’s
capital assets becomes a very complex undertaking.
The Intel Technical Manufacturing Engineering (TME)
organization is responsible for keeping this process on track,
and for doing it affordably. This group of approximately 600
employees manages Intel’s capital supply chain across fve
technology generations (Figure 1). Intel TME’s responsibility for
each generation begins with research and development and
continues on through design, manufacture, installation, use,
maintenance, and end-of-life management.
Intel TME works with hundreds of internal and external groups,
including commercial and academic research teams, tool design
and manufacturing vendors, and many others. The decision
matrix for keeping core processes on track is complex. Issues
such as technology, velocity, sustainability, and affordability
must be balanced in every decision.
COLLABORATIVE
RESEARCH
INTERNAL
RESEARCH
DEVELOPMENT MANUFACTURING END OF LIFE
Leveraging
Investments
Securing the Right
Technologies
Choosing the Right
Partners
Delivering to the Core
Extending the Value
Spending Money
Wisely
Extracting Full
Value
Figure 1. From research and development to end-of-life, the Intel Technology Manufacturing and Engineering group manages the lifecycle of Intel’s
capital manufacturing equipment to deliver affordable innovation at the speed of Moore’s Law.
2
Implementing Self-Service BI to Improve Business Decision Making
The Information Challenge: Scattered Data and Static Reports
Intel TME generates large amounts of data in managing Intel’s capital supply chain. Much of this data is scattered throughout the
organization. It resides within production applications, databases, and data warehouses at the departmental and enterprise levels,
and also within Excel spreadsheets and Microsoft Access* databases owned and managed by individual employees. In most cases,
employees use Excel and manual methods to process, update, and maintain the data, and they create presentations on an ad hoc basis
by copying and pasting Excel tables and graphs into Microsoft PowerPoint* presentations (Figure 2).
Until recently, there has been no unifed focus on standardizing
supply chain data and validating its quality across the Intel
TME organization. However, the pace of business continues to
accelerate at Intel, as it does throughout the business world.
Faster access to data and better, faster decision making is
increasingly necessary to keep pace with demands, especially
as complexity continues to climb. It is no longer tenable to
take months to deliver data and information. Better tools and
methods are needed.
Implementing Next-Generation BI
In most cases, the frst step in improving decision making is to
speed “time-to-data.” This enables individuals to devote more
of their time and effort to exploring and analyzing the available
data before their decision window closes (Figure 3). Ideally, this
results in both faster and better decisions that deliver higher
value to the business.
Other Data
Source(s)
Microsoft
SQL Server*
Individual Access
Business User, Data Analyst,
Commodity Manager, etc.
Microsoft
Excel*
Microsoft
PowerPoint*
Figure 3. The first step in improving decision making is speeding
“time-to-data,” so employees can focus more of their time and effort
on analyzing the available information to make higher quality decisions
in less total time.
Figure 2. In most cases, Intel Technology Manufacturing and Engineering employees currently use Microsoft Excel* and manual methods to process,
update, and maintain data, and they create presentations by copying and pasting Excel tables and graphs into Microsoft PowerPoint* presentations.
TIME-TO-DATA
TIME-TO-DATA
TIME-TO-DECISION
TIME-TO-DECISION
TIME-TO-
INFORMATION
TIME-TO-
INFORMATION
DECISION
DECISION
New Technologies Create New Opportunities
A number of technologies and solutions have emerged recently
that can help organizations speed time-to-data for individual
decision makers. Two new capabilities stand out as particularly
relevant.
• Real-time analytics acting on large data sets. This capability
is typically delivered by a combination of several database
technologies, such as in-memory analytics, columnar data
structures, and data compression. These technologies take
advantage of the parallel execution capabilities and large
memory capacities delivered in recent server generations.
For example, a four-socket server based on the Intel® Xeon®
processor E7 family
1
now provides up to 80 execution threads
and supports up to 2 terabytes of memory. With these
resources, even very large data sets can be held entirely in
memory and large numbers of concurrent queries can be
executed in parallel.
• Self-service BI. Historically, an end user interested in accessing
business data and creating visualizations would require end-to-
end assistance from IT staff. A new class of easy-to-use BI tools
is now available that allows typical business decision makers to
quickly access, combine, and analyze data, create presentations,
and then share the results with others.
In combination, these innovations allow IT organizations to shift
their focus. Instead of delivering reports, they have the option
of delivering data models and visualization tools to the business
and allowing users to create their own reports. Although there
are many sophisticated analytic techniques that still require
expert data analysts and developers, these self-service tools
fll a critical gap to enable more effcient and fexible delivery
of enterprise information. 3
Implementing Self-Service BI to Improve Business Decision Making
3
The Intel TME implementation of the Microsoft BI solution
stack is shown in Figure 4. The back end of the BI environment
consists of an organizational data mart built on SQL Server. The
team uses a number of built-in services to manage the data
mart, including the following:
• SQL Server Integration Services (SSIS) for deploying extract,
transform, and load (ETL) packages to configure data feeds
from external sources
• SQL Server Analysis Services (SSAS) for deploying and
managing cubes and tabular models within the data mart
• SQL Server Reporting Services (SSRS) for creating reports
• Microsoft Visual Studio* with BI Development Studio (BIDS)
and SQL Server Data Tools (SSDT) for development teams
to write BI applications, including ETL packages, data models,
reports, and dashboards
Implementing the Microsoft BI Solution Stack
Intel TME’s initial drive to improve BI came from two directions.
The lengthy timelines associated with traditional BI projects
were delaying business access to information and also placing a
heavy burden on BI development teams. A business analyst in
the Intel TME capital planning organization was engaging with
key staff to address this issue. At the same time, a software
developer in the Intel TME Systems Engineering group was
exploring technical solutions using the Microsoft BI stack and
Kimball data warehousing methodologies.
The employees driving these independent efforts connected,
providing a fortuitous combination of need and solution. Initially
the two employees focused on a single business need: measuring
supplier delivery performance. This frst project delivered
suffcient business value to justify the creation of a formal Intel
TME BI team to develop and execute a roadmap.
The Microsoft BI solution stack aligns well with Intel needs
and requirements. Most Intel TME data resides in SQL Server
databases and multi-dimensional cubes, personal Microsoft
Access databases, and Excel spreadsheets. Microsoft BI tools
make it easier to take advantage of these data sources and to
leverage existing knowledge and skillsets.
INTEL TME BI ARCHITECTURE
Tabular Mode
Multidimensional Mode
Microsoft SharePoint
Hardware
Infrastructure
Microsoft SQL Server
Analysis Services (SSAS)
Excel Services
Power View for SharePoint
(Reporting Services)
PowerPivot for SharePoint
(Analysis Services)
Microsoft
Internet Explorer
PowerPivot
for Excel
Microsoft Excel
Third Party Apps
Team BI Personal BI Organizational BI
Storage
Clustered
Virtual Hosts
Virtual
Machines
4 vCPUs, 12GB memory
(production, test, development)
16TB shared SAN
5 ProLiant BL490c G7 Blades
(2 x Intel® Xeon®processor 5600 series and 400GB memory per blade)
4vCPUs, 6GB memory
(production, test, development)
Figure 4. The Intel Technology Manufacturing and Engineering BI implementation provides users with flexible access to data, reports, and BI tools
using a Microsoft BI solution stack that includes Excel*, Internet Explorer*, SharePoint*, and an organizational data mart built on Microsoft SQL Server*.
4
Implementing Self-Service BI to Improve Business Decision Making
UPGRADING THE SOFTWARE STACK TO DELIVER HIGHER VALUE
The Intel Technology Manufacturing and Engineering BI solution stack currently includes Microsoft SQL Server* 2012,
SharePoint* 2010, Windows Server* 2008 R2, and Excel* 2010, as well as System Center 2007 for assisting infrastructure
management teams. Intel is currently in the testing and planning stages of an upgrade that will include moving to SharePoint
2013, Windows Server 2012, Excel 2013, and System Center 2012. The upgrade will most likely include a move to new blade
servers based on the latest Intel® Xeon® processor E5 family to reduce the data center space and power footprint.
There are many advantages to this updated solution stack. One of the most valuable for Intel TME is that PowerPivot and
Power View will be fully integrated across all software components, so tabular data models will be supported throughout the
BI environment, with no need for plug-ins or workarounds. This will allow the team to provide users with direct access to data
models within the Intel TME data mart to support full self-service BI functionality. As part of the move to Windows Server 2012,
the team is currently testing the latest version of Hyper-V*, and anticipates being able to reduce software licensing costs for
supporting the virtualized server environment.
For more information about Microsoft BI solutions and the Intel Xeon processor E5 family,
see the resource links at the end of this paper.
SharePoint provides the foundation for managing end-user
access to data and libraries of shared reports. End users have
the option of logging onto SharePoint with their browsers and
using server-side instances of Excel to view and interact with
data and reports. Alternatively, they can download data and
reports for local execution on their personal instances of Excel.
(Note: The Intel TME BI team currently provides access to shared
reports. Providing users with direct access to data models is a
work in progress. See the sidebar, Upgrading the Software Stack
to Deliver Higher Value.)
The Intel TME BI team manages access rights using Microsoft
Active Directory*. The security model extends into SQL Server,
which enables granular control of data access. This is important,
since the Intel TME data mart contains some of Intel’s most
sensitive intellectual property (IP).
The hardware infrastructure for the Intel TME BI implementation
consists of fve clustered, two-socket blade servers, each
confgured with two six-core Intel® Xeon® processors and 400 GB
of memory. These servers currently host nine virtual machines:
two for the production environment, two for staging, and fve
for development. All instances of SQL Server run in a single
virtual machine confgured with 4 vCPU and 6 gigabytes (GB)
of memory. All instances of SharePoint Server run in a single
virtual machine confgured with 4 vCPU and 12 GB of memory.
Infrastructure teams are currently exploring the advantages of
increasing memory to 16 GB for the SharePoint virtual machines.
The large amount of system memory is provided to support
Microsoft xVelocity In-Memory Technology (discussed below),
which provides fast data manipulation and queries by holding
all data in memory, rather than shuttling data back and forth
between memory and disk.
Ensuring Data Quality
Data quality and completeness are critical success factors for
all BI efforts. Most data in the Intel TME data mart comes from
Intel TME production applications and databases. As the Intel
TME BI team starts processing data for BI consumption, they
also perform basic auditing to ensure validity. This is especially
critical for data that will be used for measures and indicators.
Since data is consumed differently for BI than it is in other use
cases, auditing often uncovers data quality issues that would
most likely go unnoticed without these efforts. Once an issue is
identifed and a root cause is determined, a project is launched to
close the data quality gap.
5
Implementing Self-Service BI to Improve Business Decision Making
5
The Intel TME BI team uses several tools to audit data, including
those provided in SQL Server 2012. The team recently began
providing select business users with direct access to the data
mart’s staging environment, facts, and dimension tables using
Excel. They have found that business users provide valuable
insights into data quality and usability.
Unifying BI across All User Groups
The Intel TME data mart was initially deployed on SQL Server
2008 R2. However, a recent upgrade to SQL Server 2012 has
delivered important new features. Combined with capabilities
available in Excel and SharePoint, these enhancements are
key to delivering effective self-service BI and integrating it
effciently with centralized BI development efforts.
• The Microsoft BI Semantic Model (BISM) in SQL Server 2012
includes new tabular data models in SSAS, as well as more
traditional multidimensional models. These tabular models can
be accessed and consumed directly by business decision makers
using PowerPivot and Power View.
• PowerPivot allows end users to manipulate large pivot tables
and tabular data models at high speed within an Excel workbook.
High performance and scalability are enabled by Microsoft
xVelocity In-Memory Technology, which combines in-memory
analytics processing with columnar data models and advanced
data compression. PowerPivot is available as a plug-in for Excel
2010, and is supported in SharePoint 2010. Microsoft has also
integrated xVelocity In-Memory Technology into SQL Server
2012 to accelerate performance for server-side data analytics.
• Power View is a self-service reporting tool for business users
that is integrated into SharePoint 2010 and into SQL Server
2012 (as part of SSRS). Power View provides an interactive
interface for users to explore, manipulate, and present the
data located in tabular data models and pivot tables. Both
PowerPivot and Power View will be included in Excel 2013,
which the Intel TME BI team is currently testing and plans to
deploy in conjunction with SharePoint 2013.
With BISM, PowerPivot, and Power View, tabular data models
are now supported across the entire Microsoft BI solution stack,
including SQL Server, SharePoint, and Excel. This consistency
enables a unifed BI environment across all usage models.
For Intel TME, this is one of the primary advantages of the
Microsoft BI solution. It is now relatively simple and seamless
for development teams to:
• Create BISM tabular models using Visual Studio SQL Server
Data Tools (SSDT). These models are then deployed to SSAS
tabular instances to be consumed by Power View reports.
• Use the Data Analysis Expressions (DAX) language within
tabular models to quickly create and deploy both simple
and extremely complex data aggregations. Unlike SSIS, DAX
provides a very efficient way to create new measures and
indicators and quickly deploy them from development, to
testing, to production environments.
With this foundation, centralized teams have an effcient BI
development framework for developing tabular data models.
End users can potentially access, manipulate, and consume those
same data models using Excel with PowerPivot and Power View.
The result is a fexible and unifed BI environment capable of
supporting the full range of Intel TME BI requirements.
Optimizing BI for Agility and Effectiveness
Like most organizations, Intel TME has a continuum of data
needs. At one end are the “big deals” that are directly tied
to core business processes and objectives, such as vendor
assessment and procurement tracking. These processes depend
on large amounts of scattered data and have a direct and
signifcant impact on Intel’s overall success. At the other end
of the continuum are the data needs of individuals and small
groups. Improving data access for these users would not have a
major impact on core business functions but might, collectively,
have a substantial impact on overall business success.
Targeting Big Needs with a Centralized Team
The Intel TME BI Working Group was formed to drive
organizational alignment on the big deals by initiating and
managing BI deliverables and working to enable successful
integration into Intel TME operations. The team gathers
requirements from each participating department and
develops a prioritized roadmap for implementation by the
Intel TME BI team.
Seven major user groups within Intel TME are represented
within the Intel TME BI Working Group, with responsibilities
ranging from equipment development and capital planning,
to training and factory ramping, to business operations and
capital equipment resale. Most of the representatives are
“data stewards” for their organizations. These data stewards
are typically Excel power users and have a good, practical
understanding of business processes, data sources, and
information needs. A few are experts in data analytics, but
most are not. A few organizations lack anyone with the required
skillsets, in which case advocates are enlisted from other groups.
6
Implementing Self-Service BI to Improve Business Decision Making
Once high-level requirements have been gathered, projects are evaluated and prioritized by the Intel TME BI Working Group. In an
attempt to avoid the kinds of political issues that can arise in cross-organizational decision making, the group developed a standardized
process for evaluating projects based on estimated risk and reward (Figure 5). Projects with low risks and high rewards get top priority
(see the sidebar, Characteristics of a Successful BI Project).
Risk
B
u
s
i
n
e
s
s
V
a
l
u
e
Project A
Project B
Project C
Project D
Project E
Project F
Project G
Project H
Project I
Figure 5. To allocate resources in the most effective manner, the Intel
Technology Manufacturing and Engineering BI Working Group developed
a standardized process for evaluating projects based on estimated risk
and reward.
Depending on the requirements, some combination of database
administrators (DBAs), data analysts, business analysts, and
BI software developers work together on each project. These
teams have found that BI development is very different from
traditional application development. Instead of building a
functional map based on business process needs, developers
must begin with a clear understanding of data sources and then
map business needs to create measures and indicators of the
available data. The process for a typical BI project includes four
key steps.
1. Engage with business teams to develop user stories that
defne requirements, including data, measures, indicators,
and reports. This step typically consumes approximately
three-fourths of the total time and effort for each project,
and the refnement of these requirements typically extends
throughout the development process.
2. Develop ETL packages, tabular data models, and data mart
changes using Microsoft Visual Studio, SSIS, and SSAS. At
this stage of Intel TME’s BI evolution, this step often involves
accessing, auditing, and integrating new data sources.
3. Create the associated reports and dashboards using Power
View. Work with data stewards and their business teams to
test and refne the solution.
4. Release the fnalized reports to production and publish
them on SharePoint as part of the Intel POR.
CHARACTERISTICS OF A SUCCESSFUL BI PROJECT
The Intel Technology Manufacturing and Engineering BI development team has found that active support from targeted
business groups is crucial to the success of BI development projects. When assessing the viability of BI requests from
business units, proposed projects should be:
• Tied to a documented business process with a clearly envisioned end-state.
• Supported actively by business representatives during requirements defnition, data validation and business process
alignment. These phases typically account for about three fourths of the work in BI development, and it is important the
business representatives are available and responsive.
• Championed by one or more key business users to fuel post-rollout adoption and to provide some level
of training and support for novice business users.
7
Implementing Self-Service BI to Improve Business Decision Making
Extending Value with BI “Soft-Serve”
Intel TME POR reports can be published only by the Intel TME BI
team, which follows standardized procedures to maintain quality
and reliability. Decision makers can use Power View to access
these POR reports (Figure 6). They can also copy the reports into
their personal sandbox areas and adapt them in any way they
choose. These altered reports are no longer POR, but they can
be widely shared. Access is limited only by the Active Directory
groups the Intel TME BI team maintains to control access to sites,
sub-sites, and libraries. Since these sandbox reports point to the
same tabular database, the data is secured no matter who views
a specifc sandbox report.
TME Data Mart
Other Data Sources
TME Depts
BI Working Group
Prioritized
Roadmap
Analyst
Analyst
User Stories
Developer
Business User
Additional
BI Requests
Self-Service
BI Opportunity
Business User
Microsoft SharePoint
Microsoft
SQL Server
Reports
“SOFT-SERVE” BI
Figure 6. As an interim step before implementing full self-service business intelligence (BI), the Intel Technology Manufacturing and Engineering
“soft-serve” BI approach allows users to access reports and adapt them to meet their specific requirements.
Intel TME calls this BI model “soft-serve.” Unlike a full self-service
BI model, the soft-serve model does not allow end users to
access and manipulate data from the data mart, only reports.
However, even this interim soft-serve model has substantial
value, since it enables the Intel TME BI team to standardize
data, measures, indicators, reports, and visualizations, while
empowering end users to modify reports as needed to address
their unique requirements.
8
Implementing Self-Service BI to Improve Business Decision Making
The Need for Speed – Improving Velocity through
Self-Service BI
The soft-serve approach to BI helps to improve fexibility for
end users, but not enough to meet the rapidly growing business
demand within the Intel TME organization. Business decision
makers want even greater fexibility, and they don’t want to
wait on lengthy centralized projects to get the information they
want in the way they want it. At the same time, the TME BI team
needs to be able to standardize and centralize TME data and
information to ensure data quality. These competing needs place
unsustainable demands on centralized IT resources. Like IT shops
everywhere, the Intel TME BI team needs a way to standardize
and validate organizational data, while simultaneously delivering
more agile and differentiated support to end users.
To address these needs, Intel TME is working to deliver full self-
service BI to data stewards and other end users. This approach
allows users to access data models within the TME data mart
and create their own reports using Power View. They can also
download the data models into their personal sandbox areas,
after which they have complete fexibility to add or remove
data, restructure models, and create their own relationships,
calculated columns, measures, and key performance
indicators (KPIs).
2
Users can then review and refne their data models and reports
in collaboration with their own business units and make the
reports available to authorized users through SharePoint. When
a user-generated report is mature, shown to deliver high value,
and approved by the appropriate business unit, the Intel TME
BI Working Group can coordinate a Technical Review Group
(TRG) to create a more complete user story, perform data
audits, determine data gaps, and enhance the report to ensure
standardized measures, metrics, and KPIs.
The new data model and report can then be deployed within
the Intel TME data mart and become part of the POR. The cycle
can then repeat, with users continuing to extend and adapt POR
reports within their private sandboxes to address their unique
needs and then sharing the results with others.
This self-service approach to BI represents a paradigm shift in
data and information delivery. It is already impacting the way
Intel TME develops and delivers data and reports. Traditional
agile processes for software development are still valuable for
centralized solutions, but they are being adapted to incorporate
elements of this more decentralized approach. Business unit data
stewards are being integrated into the development process,
and soft-serve and self-service BI is becoming an essential and
integrated part of the overall offering.
Soft-serve and self-service BI introduce risks as well as rewards.
Without centralized control, not all data models and reports will
be high quality. Some may even provide wrong or misleading
information. However, it is Intel’s belief that the potential value
far exceeds the risk by providing an exceptionally broad base
for innovation. Hundreds of users, all with different needs,
experience, and insight into business processes, will be able to
access and analyze data faster and more effectively.
Intel TME is not the only organization within Intel that can
beneft from faster access to data and information, so Intel
is extending access to self-service BI tools throughout the
enterprise. This broader self-service BI implementation uses the
same software stack, but a different physical implementation.
3
It went live in 2011 and had over 5,000 users within one year.
By mid-March 2013, there were 13,800 active users
and the user base continues to grow.
9
Implementing Self-Service BI to Improve Business Decision Making
A Decentralized Approach to Training and Support
Effective use of self-service BI requires that users understand
the following:
• Relevant business processes
• Pertinent data sets, including how they are organized and
interrelated and how to access them
• Tools and methods for analyzing, presenting, and sharing data
It would be possible to train large user groups to use BI tools, but
business processes and relevant data vary across the company.
Because of this, training end users for BI will never be like
training end users to use a traditional business application. A
decentralized approach to BI training helps to address this issue.
PowerPoint
Self-Service BI
TRG
TME BI Roadmap
User Stories
Project Status
Ready for
TME BI?
KANBAN
TME BI Application
LIfecycle Management
Create/Update
Update
Deploy
Review
TME Dept
Microsoft
SharePoint
Yes
No
Data Steward
Data Steward
Data
Steward
Data Steward
Analyst
Developer
Business User
Data Steward
Other
Data Sources
TME
Data Mart
SELF-SERVICE AND CENTRALIZED BI INTEGRATION
Microsoft
SQL Server
Figure 7. Intel Technology Manufacturing and Engineering is moving toward full self-service business intelligence, in which data stewards and other
users can access data as well as reports and adapt both in any way they choose within their personal sandbox environments.
The Intel TME BI team is educating data stewards to use the BI
tools. Intel corporate IT takes a similar approach, training site
owners and select users throughout the business. In both cases,
those individuals then train and support people within their own
organizations. This process helps to get the right information and
training to the right people, by creating expanding “communities
of excellence” tailored to the needs of each organization.
Intel IT offers some additional training sessions on an ad hoc
basis to help users address common challenges. Users can also
attend online “offce hours” to ask specifc questions that may
be beyond the expertise of their data stewards. For now, this
strategy is working well, enabling a small team to train and
support a very large user group. As self-serve BI continues to
increase in popularity and importance at Intel, it may become
necessary to deliver more formal training to supplement the
current programs.
10
Implementing Self-Service BI to Improve Business Decision Making
Moving Forward
Success to Date
The Intel TME soft-serve and self-service BI programs are still
in a relatively early stage. The Intel TME BI team released one
centralized report to production in 2011 and fve more in 2012.
Many more are in the development queue for 2013. One example
of a successful BI deliverable is the Intel TME Forecast Progress
Indicators report, which is used to help Intel TME stakeholders
understand the current status of the capital supply chain.
Intel TME has never before had a single view showing the status
of all the tools at each supply chain stage, much less with this
level of detail. The report provides a summary page showing
the key steps—Forecast, Order, Delivery, Impact, Transit, and
Install—and provides drill down alternatives to view graphs and
explore details. Gaps can be identifed quickly and explored to
help mitigate risks and keep factory ramping schedules on track.
This information is now delivered automatically to all
stakeholders, so it is no longer necessary to create and run
manual reports. The Intel TME Forecast Progress Indicators
report and several others have reduced time-to-data to help
business decision makers monitor, evaluate, and manage key
processes, including the following:
• Supplier delivery performance
• Procurement status across the capital supply chain
• Risk factors for equipment implementation and quality
• Document readiness
In one case, the delivered report provided approximately 400
hours per quarter in productivity savings. In general, however,
quantifying value in the BI arena is challenging. Individual
decisions must be tracked to determine if they were, in fact, the
best decisions. The speed and quality of the decision must also
be tied back to the BI environment.
Because of these challenges, Intel TME is currently relying on
user perception and feedback to evaluate success. Feedback has
been positive across the majority of projects:
“ In the past, we had to pull and link all this data together
manually from the different systems. The process was so time-
intensive, we only did it once per quarter—at most. Now it takes
just an hour or two to pull and manipulate the data.”
“ This tool has facilitated an in-depth analysis at the SRC line
item level, allowing us to analyze point allocation and actual
scores on the spot. We couldn’t do this before.”
“ For years, I’ve been asking how we could measure work
profciency. Now I have an answer. Great job! ”
“ Thank you for your support in generating tailored data
acquisition pulls and metrics for measuring continuous
improvement. We’ll be happy to use these indicators!!
And more!!! ”
The majority of Intel business users also report an excellent
experience using Excel, PowerPivot, and Power View for
accessing and creating reports and dashboards. Even very large
data sets can be combined, fltered, and processed with fast
application response times and building reports is relatively
simple and intuitive. The ability to continue manipulating data
and visualizations within existing presentations has proved
valuable in creating impactful and easily customizable reports.
This feedback is valuable, but more quantitative data is needed
for effective planning. Efforts are underway to introduce these
and other success metrics in 2013. The primary focus will be on
quantifying the impact reports have on time-to-data. Once the
team has a handle on these metrics, they will begin focusing on
more complex issues, such as decision quality and the ability to
make more decisions in less time.
Future Plans
Although the BI program has delivered clear benefts to date, the
Intel TME BI team believes the biggest beneft is the groundwork
that has been laid for future BI innovation by standardizing
tools and processes for data access, analysis, and visualization.
Processes are becoming more repeatable and success more
predictable. The amount of high-quality data in the Intel TME
data mart is also growing, which will create new opportunities
and faster paths to success going forward.
BI and analytics are a work in progress at Intel, as they are for
every company. The Intel TME vision for BI is a world in which
the following criteria are satisfied:
• Every employee has 24x7x365 access to the data and
information they need to do their job, through a single portal
that is available worldwide and configured to match their
unique role in the business.
• Portal usage is intuitive, with self-serve data and no need
for training.
• Barriers to data access are minimized within the constraints
dictated by the need to protect Intel IP.
• All data accessed by decision makers comes from a single
system of record to ensure data quality and a single version of
the truth. To meet this need, Intel TME must be able to access
all required data across the Intel Technology Manufacturing
Group
4
(TMG) and the Intel TMG supplier ecosystem. This, in
turn, requires that Intel TME data should be fully aligned with
supplier data obtained through the next-generation Intel TMG
Supplier Presence Site.
• BI tools are embraced by all levels of the organization, to
the extent that individuals and teams update their business
processes to make better use of this new information source.
In time, manual sources of information should disappear
completely, because they are no longer needed.
11
Implementing Self-Service BI to Improve Business Decision Making
Additional Resources
For more information about:
• The Microsoft BI solution stack, visit:
www.microsoft.com/en-us/bi/Capabilities.aspx
• Intel® Server Products, including the Intel Xeon processor E7
family, the Intel Xeon processor E5 family, and the Intel® Xeon®
processor E3 family, visit: www.intel.com/content/www/us/
en/processors/xeon/xeon-processor-e7-family.html
It should be emphasized that these are goals, not expectations.
Existing data sets are growing and data sources are proliferating
throughout the business. Fully controlling data and information
will most likely never be a practical possibility. Nevertheless, a
vision of an ideal BI environment helps to guide long-term plans
and has helped the Intel TME BI team establish the following
goals for 2013.
• Manage centralized BI projects as a portfolio. Advances
in quantifying BI project success will help in this endeavor,
providing a better foundation for prioritizing projects and
allocating resources based on expected returns.
• Move toward trending and predictive analytics. Efforts to
date have focused primarily on tracking current activities with
reference to goals and expectations. The team has performed
a few “what if” analyses to try to predict outcomes for better
decision making and sees substantial value in moving toward
this more advanced analytics model. Efforts will continue to
evolve as the TME BI team builds a better understanding of
both technical and business opportunities and hurdles.
• Strengthen BI alignment across Intel. Collaborative efforts
are underway with other groups to define self-service BI at
Intel. As one example, the TME BI team is partnering with the
SharePoint Business Intelligence enterprise team, which is
building an Intel-wide Community of Excellence program around
SharePoint BI.
Conclusion
Evolving a global supply chain at the speed of Moore’s Law is
a complex undertaking, requiring tight coordination among
hundreds of companies and thousands of individuals. Intel TME is
using an integrated combination of centralized and self-service
BI development to provide decision makers with simpler and
faster access to the information they need to do their jobs more
effectively. Microsoft BI tools provide the software foundation
for this integration, with an organizational data mart built on SQL
Server, a managed BI portal built on SharePoint, and end-user
tools provided by Excel (with PowerPivot and Power View)—all
running on Intel® processor-based servers and laptops.
The ability to use consistent data models for both end-users
and centralized development teams has been instrumental in
providing fexible and effective support for all user groups.
Business users can access centrally developed and validated POR
reports to improve decision making for core business functions.
They also have the fexibility to adapt and extend those reports
with new data and alternative analyses, or to build entirely new
data models and reports. As information requirements continue
to grow, the Intel TME BI team believes this strategy will enable
continual improvements in data usage and decision making
throughout the organization.
Implementing Self-Service BI to Improve Business Decision Making
1
For more information on the Intel
®
Xeon
®
processor E7 family, visit the Intel web site. www.intel.com/content/www/us/en/processors/xeon/xeon-processor-e7-family.html
2
Power View reports created in Microsoft Excel* instances currently cannot be imported directly into Microsoft SharePoint* and consumed using reporting tools within SQL Server* Reporting Services, even though the PowerPivot and tabular
data models are typically the same. Technology Manufacturing and Engineering IT is currently exploring options for eliminating this gap.
3
Both the Technology Manufacturing and Engineering implementation and the broader, self-serve implementation use the Microsoft BI solution stack. However, there may be differences in product generations and version numbers.
4
The Intel Technology Manufacturing Engineering (TME) organization exists within the larger Intel Technology Manufacturing Group (TMG) , which is responsible for all aspects of Intel’s global manufacturing assets.
INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL
®
PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE,
TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. EXCEPT AS PROVIDED IN INTEL’S TERMS AND CONDITIONS OF SALE FOR SUCH
PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO SALE AND/OR USE OF INTEL
PRODUCTS INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT,
COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT. UNLESS OTHERWISE AGREED IN WRITING BY INTEL, THE INTEL PRODUCTS ARE NOT DESIGNED NOR
INTENDED FOR ANY APPLICATION IN WHICH THE FAILURE OF THE INTEL PRODUCT COULD CREATE A SITUATION WHERE PERSONAL INJURY OR DEATH MAY OCCUR.
Intel may make changes to specifcations and product descriptions at any time, without notice. Designers must not rely on the absence or characteristics of any features or instructions
marked “reserved” or “undefned.” Intel reserves these for future defnition and shall have no responsibility whatsoever for conficts or incompatibilities arising from future changes to
them. The information here is subject to change without notice. Do not fnalize a design with this information.
The products described in this document may contain design defects or errors known as errata which may cause the product to deviate from published specifcations. Current
characterized errata are available on request. Contact your local Intel sales offce or your distributor to obtain the latest specifcations and before placing your product order. Copies
of documents which have an order number and are referenced in this document, or other Intel literature, may be obtained by calling 1-800-548-4725, or by visiting Intel’s Web site
at www.intel.com.
Copyright © 2013 Intel Corporation. All rights reserved. Intel, the Intel logo, and Xeon are trademarks of Intel Corporation in the U.S. and other countries.
*Other names and brands may be claimed as the property of others. Printed in USA 0413/RS/MESH/PDF Please Recycle 328830-002US
doc_965604906.pdf