Paper

• Data-driven business insights are
available almost instantly. Complex
queries can be completed in a matter
of seconds or fractions of a second, and
the results can be funneled immediately
back into transactional applications to
improve outcomes.
A number of technology advances
have come together to make in-memory
computing a viable option. The first is the
availability of affordable memory modules.
The cost of memory has declined for years
and has reached a tipping point at which
the value of in-memory computing can
exceed the cost in many business scenar-
ios. According to Gartner, memory costs
continue to drop by roughly 32 percent
per year,
1
so the economics of in-memory
computing will continue to improve.
Businesses need more than affordable
memory to implement new in-memory
computing models. They also need produc-
tion-ready hardware and software solutions,
and both are available today. Servers based
on the Intel® Xeon® processor E7 v2 family
provide the memory capacity, execution
resources, and reliability needed to support
in-memory computing solutions at enter-
prise-scale. In-memory database software is
also available today from dozens of vendors.
The transition to this new paradigm has
begun and will continue to accelerate. The
potential benefits are already too great
to ignore for many businesses, and the
associated costs will continue to fall.
The Power of Real-Time
Business Intelligence
In-memory computing is already driving
real-world business innovation. There
are several production-ready in-memory
databases available today. A prime example
is SAP HANA,* which was developed jointly
by Intel and SAP. More than a thousand
customers across nearly 30 industries
have implemented SAP HANA and have
realized an average performance gain
of 12,380 times.
6
For example:
7
• Yodobashi, a leading Japanese retailer
calculates loyalty card points for 5 million
members. This process used to take
three days and was performed once per
month. Now it can be completed in just
two seconds. The resulting insights are
used to improve outcomes in real-time
customer engagements by providing
personalized offers based on loyalty
status and product inventories.
• Centrica, a large utility company, rolled
out smart meter technology across the
United Kingdom. It used to take three
days to analyze the data. The calculations
are now completed in seconds to improve
pricing decisions and forecast accu racy,
and to provide consumers and businesses
with customized information, services,
and products.
• T-Mobile, one of the world’s leading wire-
less providers, runs reports on 2 billion
customer records in just 5 seconds to
provide targeting marketing to 21 million
customers. The analyses are not only
much faster than the company’s previous
solution, but also much deeper. Instead of
three months of customer records, they
now analyze 24 months for better insight
into customer buying patterns.
The speed and scale of SAP HANA can be
applied to almost any data-intensive busi-
ness challenge. It is in production use for
functions as diverse as retail stocking and
supply chain optimization, global financial
closes, personalized medical treatments,
and genomic analysis. Across all of these
usage models and many more, insights are
available in seconds rather than hours or
days. New business models are possible
and customers can analyze their data far
more extensively without overloading
their computing infrastructure.
REAL-TIME, DATA-DRIVEN
INSIGHTS AT ENTERPRISE SCALE
Traditional RDBMS
Big Data Integration
(Apache Hadoop*)
OTHER ENTERPRISE
DATA SOURCES
Intel® Xeon®
Processor E7 v2
Family
A SINGLE HIGH-SPEED DATABASE FOR
TRANSACTIONS AND ANALYTICS
Insights available in seconds
IN-MEMORY COMPUTING: A GAME-CHANGING PARADIGM SHIFT
In-Memory
Database
Software
+
Real-Time
Data Replication
Real-Time
Ad Hoc Queries
Figure 2. With in-memory computing, traditional latencies are greatly reduced to enable real-time, data-driven decision making based on all relevant
data sets.
“ It is now possible to store a primary
database in silicon-based main
memory, resulting in an orders-
of-magnitude improvement in
performance and enabling the
development of completely
new applications."
5
— The Global Information Technology
Re po r t 2012
4
Changing the Way Businesses Compute…and Compete