Peer-to-Peer

Back-up at Otis

Otis India roped in CA to implement its Enterprise Back-up Management System. Aishwarya Ramani has the details

"While selecting a back-up system, one needs to focus on minimising the back-up window and improving the performance of the server" - V Subramaniam CIO, Otis India

Otis Elevator India’s IT set-up consisted of distributed IT systems, and it needed to consolidate its back-up requirements. The company had deployed multiple operating systems at its locations across the country. It had been using Unix as the server operating system, and SQL Server 2000 as its database running on servers that were scattered across locations. The company needed a unified data back-up solution across the organisation.

To consolidate its disparate data, Otis decided to use CA’s BrightStor, an integrated back-up and recovery solution. BrightStor has a Java-based GUI. Its back- end stores a process log. This product has an inbuilt back-end database known as Very Large Database. Users of BrightStor are free to use this or SQL Server.

This project was executed in 2003, prior to which they were using the inbuilt back-up mechanism that comes with Windows Server. This permitted Otis to take back-ups from individual servers, but the information had to be combined manually. This just didn’t cut it when Otis wanted to consolidate its servers. Moreover, the earlier system led to problems such as errors during back-up, data losses, and inefficiencies in managing the back-up process. Apart from this, the technology team had to travel to remote locations to solve problems related to data management. All of this prompted Otis to consolidate its servers and deploy a full-fledged back-up management system. “We call it the Enterprise Data Protection System,” says V Subramaniam, Chief Information Officer of Otis India.

The basic requirement that Otis was looking for was a system that effectively created a back-up environment and provided protection. Subramaniam was clear that the back-up system that it would implement would have to be secure as well as easy and quick to restore from. (BrightStor offers its users features such as a command line interface, and also integrated anti-virus capabilities, a list of pre-defined reports including real-time back-up and restore, and a statistics recovery report.)

A principal consideration was the back-up window or the time taken to back-up the system. Says Subramaniam, “While selecting a back-up system, one needs to focus on minimising the back-up window and improving the performance of the server.”

One of the reasons for selecting BrightStor was that it offered data protection for distributed servers, applications, as well as clients for multiple environments, including Windows, Unix, Linux and NetWare.

The storage process at Otis India is divided into three stages: back-up, store, and make available. After a back-up is taken, the tapes are stored in a fire-proof cabinet at an offsite location 30 km from Mumbai. When asked about prevention of loss due to disasters from calamities (think floods), Subramaniam said, “Our data store is well above ground level, so floods do not bother us overmuch.”

Back-ups are taken at all four regional offices—Delhi, Mumbai, Kolkata and Bangalore. BrightStor has also been implemented at the company’s Bangalore factory. Data back-ups from offices across India are brought to Mumbai. This exercise takes place every alternate day. Initially all the data was sent to the headquarters at Mumbai. Today, each location outside Mumbai is divided into sub-locations. Back-ups are first sent to these sub-locations before being sent to Mumbai.

Another aspect that the people at Otis were particular about was acknowledgements from people at offsite locations, who have to acknowledge receipt of the back-up tapes; this is mandatory.

The deployment was executed in part to adhere to regulations. The company’s back-up policy defines what it will back-up, how the back-up will be done, the frequency of the back-up, and how it is going to store this data.

The back-up system at Otis is scheduled to run everyday at a specified time which is decided by company policy. The back-up of the messages takes place online. After the back-up is complete, the system selected by Otis writes a log of the time taken to complete the process and notes errors during back-up, if any.

“This data has to be available for restoration whenever required,” says Subramaniam. Restoration can be complete or selective. “Sometimes an employee may delete an e-mail message that is of prime importance. In this case selective restoration from the message back-up will be done.”

Otis felt that going for a big-bang implementation was appropriate as all its offices needed to go live at the same time. This allowed the company to take back-ups of all its centres at one-go. The company first implemented BrightStor at its data centre in Mumbai. Explains Subramaniam, “While we implemented simultaneously across the other offices, we started with the data centre in Mumbai as it has the gateway, primary file server and authentication systems.”

Replication is also an important aspect of Otis’ back-up plans. According to Subramaniam, “The replication process is conducted online as well as offline. The applications are replicated offline whereas online replication is used for taking back-ups of email messages.” This replication mechanism is built into the solution provided by CA.