Fill It, Shut It, Retrieve It: How Hero Honda Topped Its Storage Strategy

New Delhi: | Updated: Oct 29 2002, 05:30am hrs
Auto major Hero Honda first felt the need for a storage solution when its database size started growing at 16 gigabytes per month and its existing infrastructure did not seem safe enough to carry the load.

Hero Hondas data on the serial storage architecture (SSA) box, which is an open standard for high-speed access was vulnerable to power failures and disk crashes. The other problems included the high back-up window, because of which the company had to take a shutdown of 8 hours for offline weekly backup and 12 hours with the current load. The peak load used to grow up to 80 per cent of total resources because of input-output throughput bottlenecks at month ends. Besides, the company needed a disaster recovery capable solution and decided to implement a disaster recovery site by the first quarter of 2003.

Hero Honda selected IBMs enterprise storage server (ESS) to meet its storage requirements. According to Hero Honda chief information officer S R Balasubramanian, the company invested around Rs 1.7 crore in the total solution including hardware, software and services. Hero Honda has bought 2 terabytes of storage space which, according to internal estimates, will suffice only till June 2003.

Implementation of the new solutions has benefited the organisation in terms of work efficiency, optimisation of resources and better response to end customers. Better servicing of customer orders has led to an increase in sales. Our computing infrastructure has been reliable and has succeeded in providing us with an uninterrupted work environment, said Mr Balasubramanian.

The implementation of storage server ensured data safety, the backup window is reduced by using flash copy functionality (shutdown time required is 30 minutes for offline backup as compared to 12 hours under the old arrangement). Flash copy function provides online backup with no extra load on the system during. Since the system has a mirrored flash memory, throughput bottlenecks were reduced by about 15 per cent.

The company claims to upgrade its SAP R/3 systems in 18 hours. Mr Balasubramaniam added that the reduction in the upgrade window was due to extra disk space, fall in input-output load and shorter backup window. Moreover, the company has saved in terms of further upgrade of memory and processor because of better response on the system.

The configuration process was simple since RAID-5 was pre-configured, we divided the RAID into logical disks of 16 gigabytes each, says Mr Balasubramaniam. Installation of SCSI adapters and their drivers on all the hosts, took only one day of shutdown while assigning the disks to nodes could be done on the fly, he added.

Migration of data from the existing system to the new storage server also took one day of shutdown, which was just disk to disk copy. After migrating the data some settings related to SAP/Oracle environment were required to be completed.

Since we were using TSM (Tivoli Storage Manager) and TDP (Tivoli Data Protection Agent) for SAP R/3, we configured TDP for storage server for R/3 and implemented flash-copy, after thorough testing of the same on R/3 development system. Since this configuration included a lot of testing of procedures like recovery and restoration of data, it took a lot of time, said Mr Balasubramanian.

The TDP was finally configured to take automated backup on tapes. Time schedule had been specified and the tape library LTO 3583 takes care of the rest, he added.