Oracle Exadata

1

O R A C L E E X A DATA C A S E S T U DY
Copyright © 2011, Oracle. All rights reserved. Oracle and Exadata are registered trademarks of
Oracle and/or its affiliates. Other names may be trademarks of their respective owners.
Oracle Exadata Achieves Fivefold Performance
Increase for Critical Product Research Platform
EXECUTIVE SUMMARY
A modern version of the traditional laboratory notebook, the electronic laboratory
notebook, or ELN, is becoming a critical asset in the pharmaceutical industry, giving
scientists new power to collaborate and share experimental data with colleagues from
various research divisions. Merck & Co. embraced the CambridgeSoft ELN technology
as an early adopter in 2002, and the application was quickly accepted by Merck’s global
research teams. As a result, the volume of ELN-documented experiments and number
of users exploded. Today, with over 4,000 users, the Merck ELN is one of, if not the
largest deployed ELN system in the pharmaceutical industry worldwide.
With the rapid growth of users and data volume, Merck and CambridgeSoft worked
diligently to keep pace with the expansion and simultaneously maintain performance.
As a result, changes were made to the application itself as well as to the infrastructure.
Merck’s planning made it was clear that rapid user growth was expected to continue and
correspondingly, the size of the database would also continue growing very rapidly. This
growth was expected to outpace the application of conventional approaches to maintain
acceptable performance. In addition to CambridgeSoft’s engineering enhancements and
traditional infrastructure changes, Merck decided to explore non-traditional options to
provide a long-term solution.
Merck found a solution in Oracle Exadata Database Machine, a new computing platform
that takes advantage of innovations in smart storage to radically boost system speed and
cost performance. The company rolled out the Exadata solution in late 2010, using Oracle
GoldenGate data-synching technology to make the switch in a single weekend, as
opposed to taking weeks for migration using conventional software. Essentially, there
was no business downtime.
Following the move, ELN performance surged, with Merck conservatively estimating a
fivefold increase in query response speed as well as substantially faster user login and
navigation performance, even as the company simultaneously expanded the user base
by 30%. Other improvements have included lower system resource utilization, optimized
storage space, and workload consolidation.
All told, Merck is expected to realize benefits totaling an estimated $8 million from its
Oracle Exadata investment.
Merck & Co.
Merck is a global research-driven
pharmaceutical company. Merck discovers,
develops, manufactures and markets a
broad range of innovative products to
improve human and animal health, directly
and through its joint ventures.
CambridgeSoft is a leading provider
of software and services for discovery,
analysis and collaboration to life sciences
and chemical industries, academia and
government. Products include Cloud,
Enterprise, Workgroup and Desktop
versions of Chem & Bio Office, including
Chem & Bio Draw and E-Notebook,
providing knowledge management,
chemical and biological informatics,
and scientific database solutions.
Founded: 1891
Headquarters: Whitehouse Station, NJ
Revenue: $27.4 billion (2009)

2
ORACLE PRODUCTS
• Oracle Exadata Database Machine
• Oracle GoldenGate
• Oracle Database 11g
KEY BENEFITS
• 159% ROI over 3 years
• Increased speed of ELN searches
by 5 times
• Total benefits of $8M over 3 years
• Reduced memory utilization by 40%,
reducing the need for further
immediate investment
• Reduced CPU utilization to 10%,
providing significant capacity
for growth
• Doubled the size of the user base
without impact to performance
• Increased scientist efficiency
“We really knocked it out of the park.The chemistry
group, which is one of the biggest users of E-lab
notebook, is very happy with what we delivered.”
M E R C K
Beyond the direct infrastructure and productivity savings, the exponential jump in ELN
performance enabled by Exadata is empowering scientists to quickly access experimental
data and collaborate more effectively with other research teams — a benefit that holds
tremendous potential in terms of future product development, innovation, and revenue.
With the increase in system responsiveness, Merck’s confidence in the system’s
scalability improved, as did the user experience. Scientists now have a tool that keeps
up with their way of doing business.
BACKGROUND
One of the world’s largest pharmaceutical companies, Merck & Co. discovers, develops,
manufactures and markets a broad range of innovative products to improve human and
animal health, directly and through its joint ventures. New Jersey-based Merck operates
in 140 countries1
and in 2009 significantly expanded its market share with the acquisition
of Schering-Plough, the maker of Claritin, Coppertone, and many other brands.
At the heart of Merck’s operations are the thousands of researchers and scientists who
invent, develop and test new products in a network of laboratories that span a wide
geography. Collaboration is essential to the success of these research teams: scientists
are more productive when they can quickly access the experimental results and techniques
of their colleagues. Indeed, increasing collaboration is a major reason why pharmaceutical
companies have adopted electronic laboratory notebooks, or ELNs — a secure, highly
searchable form of the traditional paper-based lab notebooks.
Merck made the move to online lab notebooks in 2002 when it rolled out CambridgeSoft
E-Notebook, a leading ELN solution. Today, ELN is a strategic asset for the company and
the primary means for capturing the intellectual property created by Merck’s scientists.
Scientists document the discoveries and conclusions of their experiments in the online
notebooks and then search the database for information to help them plan the next phase
of their experiments.
For example, a chemist working on a promising new drug compound will search the
ELN database to learn if anyone else had made a similar or related compound — and then
use that information to plan the next stages of his experiment. Searching collective
information on experiments enables Merck scientists to be more productive and leads
to greater innovation.
The growing use of ELN, however, began to put strains on system capacity and
performance. Between 2003 and 2010, the user base doubled and the size of Merck’s
ELN database grew to more than 4 terabytes.
1
From Merck corporate Website

3
Figure 1: Active Users of Merck’s ELN Application
Figure 2: Number of Experiments Documented in ELN Application
With system usage and database size soaring, information searches began to slow. “As
more and more people plugged into the ELN platform, more and more people wanted to
use it,” said Merck. “That meant that the platform’s performance got progressively worse
over time as we struggled to keep up with infrastructure enhancements. CambridgeSoft
also worked with us to streamline the search process, which yielded tangible
improvements.”
Still, slowing performance hurt researchers’ productivity, since scientists were hindered
from taking full advantage of the system to assist with research.
Merck’s initial attempts at fixing the problem with conventional methods — including adding
hardware and memory and upgrading CPUs — met with limited success as the expanding
user base, experiments and searches continued to outpace efforts to upgrade system
hardware. Consequently, in 2010, Merck resolved to permanently solve its ELN performance
issues by embracing the next generation of database and storage technologies.
ELN Active Users (>=1 login/month)
Feb ‘11
Jun ‘10
Oct ‘09
Feb ‘09
Jun ‘08
Oct ‘07
Feb ‘07
Jun ‘06
Oct ‘05
Feb ‘05
Jun ‘04
Oct ‘03
Total ELN Experiments Created
Feb ‘11
Jun ‘10
Oct ‘09
Feb ‘09
Jun ‘08
Oct ‘07
Feb ‘07
Jun ‘06
Oct ‘05
Feb ‘05
Jun ‘04
Oct ‘03
ELN Active Users (>=1 login/month)
Feb ‘11
Jun ‘10
Oct ‘09
Feb ‘09
Jun ‘08
Oct ‘07
Feb ‘07
Jun ‘06
Oct ‘05
Feb ‘05
Jun ‘04
Oct ‘03
Total ELN Experiments Created
Feb ‘11
Jun ‘10
Oct ‘09
Feb ‘09
Jun ‘08
Oct ‘07
Feb ‘07
Jun ‘06
Oct ‘05
Feb ‘05
Jun ‘04
Oct ‘03

4
THE ORACLE EXADATA SOLUTION
Oracle Exadata emerged as the logical alternative to conventional database technologies,
offering significantly faster performance for both data warehousing and online transaction
processing (OLTP) applications, including electronic lab notebooks. Oracle Exadata combines
servers, storage, networking, and software in an integrated platform that is massively
scalable, secure, and redundant.
Following a successful pilot, which included CambridgeSoft’s participation in the Exadata
optimization, Merck deployed Oracle Exadata in late 2010 as the standard platform for its
ELN solution. Because the switch to Exadata represented a major system change involving
both new hardware and new Oracle software, Merck was anticipating more than a week
of downtime for the migration. But Merck minimized downtime to a single weekend by
deploying Oracle GoldenGate to keep the old and new environments synchronized during
the transition.
“We were literally able to flip the switch to launch the new platform,” said Merck.
“Scientists left Friday afternoon and returned Monday morning to a new system that
was up and running with all the old data. There was no negative impact to the scientists.”
Moreover, Merck wasn’t required to make any changes to the ELN application itself to
run on Oracle Exadata.
OPERATIONAL AND STRATEGIC BENEFITS
In the months following the switch to Oracle Exadata, Merck has seen a host of
operational and financial benefits ranging from dramatically faster system performance
to millions of dollars in avoided solution-replacement costs. From a strategic perspective,
the move to Exadata enables Merck’s scientists to more fully utilize the ELN platform,
boosting their research capabilities and potentially speeding product innovations.
Fivefold Performance Increase
With Oracle Exadata, Merck improved ELN search speeds by at least four-to-five times,
easily meeting the IT group’s service level agreement (SLA) target of 30 seconds for
most searches. Timeouts were eliminated. “Overall, we realized about a 5x improvement
in performance, even with complex searches,” said Merck.
As shown in Table 1, the move to Exadata led to enhancements across a range of ELN
performance measures — from faster user logins and navigation, to speedier text and
structured-data searches. Merck’s IT team calls the fivefold speed boost a conservative
estimate because many of the queries in the legacy environment were terminated
before returning results. Other estimates place the performance boost at nearly 10 times
(in a controlled environment).
HOW EXADATA MAXIMIZES
PERFORMANCE
• Exadata Smart Scan. The smart
storage software in Exadata offloads
data-intensive query processing
from Oracle Database 11g servers
to Exadata’s storage layer for parallel
data processing. Because there’s
less data moving through the higher-
bandwidth connections, you get dra-
matically improved performance as
well as concurrency for simple and
complex data warehousing queries.
• Exadata Smart Flash Cache.
With more than 5 terabytes of flash
memory per full rack, Oracle Exadata
intelligently caches “hot” data and
assigns the rest to disk storage,
giving organizations the speed of
flash with the cost-effectiveness of
disk storage. Exadata Smart Flash
Cache can process up to 1.5 million
random I/O operations per second
and scan up to 50 GB of data per
second to deliver ultra-high perfor-
mance for OLTP applications.

5
“We really knocked it out of the park,” said Merck. “Our chemistry group, which is one
of the biggest users of E-lab notebook, is very happy with what we delivered.” Merck
attributes the faster search performance to two technology advancements found in the
new Oracle platform: Exadata Smart Scan, which filters out data that is not relevant to a
query, and Exadata Smart Flash Cache, which improves performance for OLTP applications.
Lower Utilization Rate
The database efficiency improvements driven by Exadata — which offloads a large portion
of data processing to the storage layer — has had a positive impact on system utilization
rates at Merck: Memory utilization is now running around at 50–60% of capacity and CPU
utilization at around 10–15% — a significant reduction in both cases.
More Productive Scientific Research
The move to Exadata means scientists are more likely to probe every promising idea.
“Combined with the CambridgeSoft ELN, Oracle Exadata allows scientists to execute more
sophisticated searches and get faster access to information they previously weren’t able
to easily access.” said Merck. This also means that there are no concerns about adding
thousands more users to the CambridgeSoft E-Notebook platform as is currently planned.
ESTIMATED $8 MILLION IN TOTAL BENEFITS
Figure 3 illustrates the total benefits of Merck’s Oracle Exadata investment over three
years. Additional savings will come from writing off Merck’s legacy database servers and
storage equipment, and from productivity improvements related to faster searches and
research collaboration.
Table 1: Performance Increase Achieved by Merck with Oracle Exadata
Before:
Average time
(in seconds)
After:
Average time
(in seconds)
Average
Performance
Improvement
Text Search (27 Steps) 99.8 23.2 4.3x
Structured Search (43 Steps) 74.2 14.0 5.3x
Structured Search Combined
withText Search (15 Steps)
232.8 40.0 5.8x

6
Copyright © 2011, Oracle. All rights reserved. Oracle and Exadata are registered trademarks of
Oracle and/or its affiliates. Other names may be trademarks of their respective owners.
Total Capex
Avoided
$5M
Productivity
Improvement
Year 1
$1M
Productivity
Improvement
Year 2
$1M
Productivity
Improvement
Year 3
$1M
Total
Beneﬁts
$8M
CONCLUSION
Facing an increasingly globalized and competitive marketplace, pharmaceutical companies
are being pressured to develop and launch products faster than ever before. That new
reality has put a premium on laboratory research efficiency and productivity. More than
ever, companies are seeking ways to accelerate the experimental process, collaborate
more efficiently, and make rapid decisions regarding research projects.
One way that Merck has responded to the challenge is by deploying Oracle Exadata to
transform the data-processing backbone of its ELN application. As a result, Merck scientists
are finding and sharing research information five times faster on average, and they are
now accessing valuable experimental data that was previously inaccessible.
The business potential unleashed by Merck’s ultra-fast ELN is enormous. Today, scientists
are significantly more productive and connected to the research of their colleagues across
the company, which is strengthening collaboration and building the foundation for rapid,
market-beating innovations.
Figure 3: Benefits by Category—Three-Year ViewABOUTTHIS CASE STUDY
Research and analysis for this study
was conducted by Mainstay Partners,
an independent consulting firm,
drawing from interviews with Merck
employees, review of planning
documents and searches of industry
literature. ROI calculations use
industry-standard assumptions
regarding the time value of money.
Mainstay Partners is the leading
provider of independent value
assessment and IT strategy services.
Our clients include Oracle, SAP,
Cisco, Lexmark, Microsoft, Motorola,
and HP. For more information,
please visit www.mainstaypartners.
net. Information contained in this
case study has been obtained from
sources considered reliable but is not
warranted by Mainstay Partners.
901 Mariners Island Blvd, Ste.105
San Mateo, California 94404-1592
650.638.0575 ph 650.638.0578 fx
www.mainstaypartners.net

Oracle Exadata

More Related Content

What's hot (6)

Similar to Oracle Exadata (20)

More from Mainstay (20)

Recently uploaded (20)

Oracle Exadata