SPECIAL MENTIONS

Continued from page: 5

Friday, June 03, 2005

Sankraman, Mahindra & Mahindra

Print

Comment

Email

Digg

Del.icio.us

Reddit

Twitter

Project Head: Krishna H Nabar, Head – Business Solutions, Corporate IT 
Location: Mumbai (Data Center) catering to around 50 locations across the country

One of the first industries in India to go for large-scale computerization was the manufacturing industry. Companies, such as Hero Honda and Maruti had gone through complex ERP implementations much before such terms even became industry buzzwords. Mahindra & Mahindra, an automobile manufacturer, also implemented SAP R/3 way back in 1997. The communications infrastructure, being poor in India at that time, forced Mahindra & Mahindra to go for a distributed implementation with SAP servers spread across eight locations all across the country. The distributed architecture brought its set of problems along; maintenance costs were high, data consolidation was cumbersome and time consuming.

Now, with the much improved communication infrastructure, and bandwidth prices having been slashed significantly, a solution to the above problems lied in a central SAP implementation which will cater to all Mahindra & Mahindra locations. Mahindra & Mahindra decided to do a fresh implementation of SAP R/3 4.7 on a single server (8 CPU, IBM P570 in a fail over cluster configuration with another 4 CPU IBM server) and then to merge data from all SAP R/3 4.6B servers into that server. As simple as it may sound, data migration from multiple servers is a complex task. It requires data collation, which should then be examined to remove redundancy and duplicity. After this, a single meaningful data instance should be created for the central server.

The implementation, which also includes a disaster recovery site, was spread across a time frame of nine months, starting from July 2004. The expected life of the implementation is five years. The implementation, while reducing administrative tasks, also brings direct business benefits by providing better performance and features than the previous implementation.

Page(s)   1  2  3  4  5  6