You are on page 1of 2

Infosys

Case Study

Solution
Case Study




Company: Infosys Limited Infosys migrates from SQL server 2012 to SQL 2016

Website: https://www.infosys.com

Company Size: 85 sales & marketing
offices, 114 development centers &
199,000+ employees globally In Infosys technology ecosystem, SQL server 2012 was being used for its Database
Country or Region: India, clients over 50 and Analytics needs. From the core Online Transaction Processing (OLTP) systems,
countries data was being synchronized onto the enterprise Data Warehouse systems which
Industry: IT, Business Consulting & in-turn would refreshes all the Analytics dashboards and reporting systems. This
Outsourcing data synchronization and staging of data on the Data Warehouse instances
normally took a day. Therefore, the Analytics and Reporting data on the
Company Profile dashboards/reports was a day old. Infosys started facing a clear demand from
Infosys is a global leader in technology services business for Real Time Analytics for informed, real-time decision making. The need
and consulting and enables clients in more than for a scalable Database platform which would provide advanced real-time
50 countries to create and execute strategies analytics was exigent.
for their digital transformation. From
engineering to application development, Let us dig a little deeper to understand how Microsofts SQL server 2016 has
knowledge management and business process transformed Analytics capabilities at Infosys.
management, Infosys helps its clients find the
right problems to solve, and to solve these
effectively. Infosyss team of 199,000+
innovators, across the globe, is differentiated
by the imagination, knowledge and experience,
across industries and technologies, that they
bring to every project they undertake.
In 1981, seven engineers started Infosys
Limited with just US$250. From the beginning,
the company was founded on the principle of
building and implementing great ideas that
drive progress for clients and enhance lives
through enterprise solutions. For over three
decades, Infosys has been a company focused
on bringing to life great ideas and enterprise
Business Needs With large ETL packages typically
running once a day, this Analytical data
solutions that drive progress for our clients.
Infosys has been using SQL server 2012 had a time lag of one day. Businesses
Infosys recognizes the importance of nurturing
and SQL Server AlwaysOn, which demanded gaining real-time insights
relationships that reflect our culture of
ensured reporting on transactional data into their data to take crucial business
unwavering ethics and mutual respect. Itll
to be near real-time. But, when it came decisions based on time.
come as no surprise, then, that 97.5 percent
to Analytical reporting, Infosys was
(Q2 FY 17) of Infosyss revenues come from
using the traditional process of Fundamentally, analytical reporting had
existing clients.
synchronization of data from Core OLTP to be real-time, this was the need of the
Infosys has a growing global presence with
systems to enterprise Data warehouse hour. Infosys was exploring various
more than 199,000+ employees. Globally, we
systems. Once the data were options to grapple with this business
have 85 sales and marketing offices and 114
transformed, cleansed, and de- need. Solutions that could aggregate
development centers as at March 31, 2016.
normalized canned reports were being and compute queries to be run directly

generated for business analytics use. on OLTP system by avoiding the entire

For more information about the case study,


kindly reach to

Shailendra.Chauhan@microsoft.com


transformation process and running 3 data centres, 12 nodes, 16 instances, be run on operational workloads using
aggregated and complex queries directly 130 discs, sizing around 90 terra bytes Columnstore indexes with no impact to
on Data Warehouse systems were being across OLTP, AG reporting and AG DR the performance of the operational
sought out. The trouble was, analytics instances . This is also the first database. With analysis happening
systems could not be built on core Distributed Availability Group (DAG) in directly on operational data, Infosys will
systems because that would lead to APAC. Various Microsoft teams be able to avoid the complexity, cost
major performance impacts something including the account management and data latency issues that accompany
the organization could not afford. team, consulting team, SQLCAT, Field the traditional methods of setting-up a
and engineering teams, worked with separate Data Warehouse instance for
What Infosys needed was an analytics Infosys at every step in the migration carrying out data analysis and reporting.
system that could provide real time data process. Right from planning,
for business analytics while not estimating infrastructure requirements, SQL Server R services: With the release
compromising on performance or carrying out the architecture, building of R Services (In-Database) along with
security issues of core systems. The servers and finally migrating this SQL server 2016, a platform has been
search ended with Infosys deciding to environment, Microsofts partnership provided for developing and deploying
migrate from SQL Server 2012 to SQL was critical in the successful intelligent applications that help
Server 2016. implementation of this missions critical business users uncover new insights.
goal. Using R Services, Infosys is able to
Solution create models to generate predictions
Benefits using the SQL server data.
SQL Server 2016 was released in June
2016. Infosys carried out the migration of Real time Reporting: The We had deployed typical architecture
the Database server to SQL 2016 with a implementation of the DAG feature of of separate clusters for OLTP and
very aggressive target of two and a half SQL server 2016 has ensured High Enterprise Data Warehouse (EDW) for
months within the product release. The Availability (HA) and disaster recovery. our mission critical transactional and
first phase of implementation was With data being available real-time 24 reporting workloads on SQL Server
carried to leverage largely on SQL by 7 from the OLTP systems to the Data 2012. Achieving real time analytics and
Servers Distributed AlwaysOn feature. Warehouse systems, an ecosystem for reporting on this architecture was not
This has enabled data to be available real-time Analytics and Reporting is feasible as we were dependent on ETLs
real-time from OLTP systems onto made available for the business teams that introduced latency. With the
enterprise Data Warehouse systems. This to leverage. release of SQL Server 2016, we
has also helped Infosys drop the innovatively architected our OLTP and
execution of Extract Transform Load Performance: The turn-around time for EDW clusters with DAG, which supports
(ETL) packages on the Data Warehouse data to be synchronized, transformed, AGs across Windows clusters. This gave
instance since data is now available real- and staged onto the Data Warehouse us best of both worlds segregated
time. instance, has been drastically cut-down clusters for OLTP and EDW along with
from one day to 1 to 2 seconds. The real time data for analytical reporting,
As part of the second phase of this migration will also eliminate the need to says Amit Kumar Tikkiwal, Principal
migration process, Infosys intends to run ETL packages for data Consultant, Infosys Limited.
leverage on in-memory and Columnstore synchronization and staging, leading to
Index Features of SQL Server 2016. This efficiency and performance boosts.
will enable real-time Analytics data to be
available on the OLTP systems without Operational Analytics and Cost
any performance impact. Effectiveness: Infosys intends to
leverage In-memory and Columnstore
With a very strong partnership with the indexes of SQL Server 2016 in the
Microsoft team, Infosys met its second phase of the migration plan. This
aggressive two and a half months will allow real time Analytics without
timeline, and created the OLTP cluster - impacting the performance of the OLTP
the largest SQL Server 2016 cluster in systems. These features of SQL Server
India with cluster nodes spanning 2 cities, 2016, ensure that Analytics queries can

This case study is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS SUMMARY.

Document published September 2015

You might also like