The Hindu Business Line : Distance makes heart fonder, data safer

Distance can mean different things to different people, but in the context of data security, the more distance between your data back-ups, the better.

IN today's competitive business environment, companies are being forced to work 24 hours a day, 7 days a week, 365 days a year. Doing business at this `365x24x7' pace generates a lot of data, and hence companies are going in for data storage systems that consolidate the huge volume of information they need to store and use.

Fine so far, but what happens when power outages, natural disasters, and the like hit these consolidated data storage systems, exposing entire databases to disruption? The obvious answer is creating data back-ups for these consolidated data storage systems.

The events of 9/11and the power outages in Europe and the US have done much to create awareness on the importance of backing up one's data so that it can be recovered, and operations quickly restored, in the event of a sudden disruptive occurrence. This is especially crucial for companies such as banks and financial institutions that maintain vast databanks of confidential customer information.

Hence, data recovery or DR systems are emerging an important component of the data storage function.

Estimates put the global storage business at $35 billion. In India, the size of the SAN, NAS (network attaches systems), DAS (direct attached servers) and DR businesses together is put at $100 million.

A jolt of awareness

According to P.P. Subramanian, Country Manager, Hitachi Data Systems, many companies that implemented consolidation last year have been going in for DR systems this year.

"Incidents such as 9/11 and the US and European power outages have shaken people's risk-taking abilities. Another fallout of 9/11 was that big corporations started moving out of noticeable buildings and areas," he says.

HDS, which set up shop in India in April 2001, began business in October 2001. "We made a slow start as there was little awareness on the need for implementing disaster-recovery systems," he says.

The company transacted business worth $10 million till October 2002 and added another $10 million during October 2002-March 2003. It is targeting business of over $20 million during the April 2003-March 2004 period.

According to Subramanian, the Government, and in particular large PSUs, have emerged as major buyers. "This year has been good, and the strong trend should carry into the next year as well. We are currently at the top of the pyramid and should move into the high-volume middle of the pyramid next year."

According to Ponnanna Uthappa, Country Manager, Allied Digital, the DR market across the world has matured as people have started realising its importance. He estimates that about 35 per cent of the entire IT market is accounted for by core computing infrastructure. And, about a third of this would be accounted for by DR systems.

Ponnanna says the Indian market is in the `initial' stages, and only the big companies have implemented full-fledged DR systems — not least because of the huge cost involved.

Distance makes a difference

Even as disaster recovery is increasingly being implemented as an integral part of companies' operational infrastructure, the concept is itself evolving.

Talking to eWorld on DR implementation abroad, Hubert M. Yoshida, Vice-President and Chief Technology Officer of Hitachi Data Systems Corporation, says implementation of disaster recovery systems has now become mandatory for financial organisations in several countries.

He says that despite the fact that 9/11 hit at the financial nerve-centre of the US, the impact was not as disastrous since many companies had backed up their systems in neighbouring New Jersey.

However, thinking ahead and taking into account the likelihood of a more widespread impact the next time round, US regulatory bodies such as the SEC and the OCC have issued DR implementation guidelines that address these larger concerns.

First, the guidelines emphasise the need for setting up `out-of-region' back-up sites to cover the threat from a "black-hole disaster," which impacts the hardware, software, infrastructure and even the personnel of a company.

The regulations also specify that there should be at least three DR sites so that the secondary back-up is not exposed, should the primary system shut down.

The idea behind having back-up sites in different regions is to ensure that it is supported and serviced by a different telecommunications provider, water and power source with a separate pool of personnel as well. In effect, the guidelines require companies to set up a separate infrastructure at the remote site.

This leads to the concept of `distance recovery' that specifies that the power grid and the entire associated infrastructure is separate.

"Distance has now become a requirement in disaster recovery," he says. Consequently DR, which has traditionally been assessed on two metrics — the `time object' called RTO and the `point object' called RPO — can now be assessed on a third metric which relates to `distance recovery.'

The distance dimension, in turn, raises the question of synchronous and asynchronous DR. Explaining, he says that in synchronous DR, data transfer from the primary DR site to the secondary and tertiary sites is accompanied by an acknowledgment of receipt at the other end. However, the distance element places limitations on synchronous transfers, which are possible between sites separated by a distance of about 60 km.

Asynchronous DR happens when data sent from one site and received at a second is collected in a buffer and then audited to see that it is in order. Such transfers can happen between sites located in different parts of the globe. Data so transferred is often jumbled at the second site. It is in order to get around this problem that Hitachi Data Systems has developed the technology to time-stamp data transfers.

Asynchronous transfers can also be facilitated through multi-hops, which are basically a series of synchronous transfers. However, since such a mode of data transfers would require replication of much infrastructure, it is a very expensive option for companies.

Yoshida says Hitachi Data Systems encourages its customers to evaluate and categorise the importance of data, based on which they can decide whether or not to replicate it.

This has given rise to a totally new trend in DR called Data Life-Cycle Management, which deals with weighing the cost of storage to the value of the data to the company.

To bring down the cost of DR, HDS also suggests categorising data based on its importance and accordingly storing it at different levels.

The most important and crucial data is given higher storage or lightning systems that store the data in a secure environment supported by up to 64 processors. Intermediate storage is ideal where data allows the opening of a maintenance window. The data stored affects a smaller number of people and the storage system has just two processors. The lower storage is ideal for data with the least importance.

anna@thehindu.co.in

Picture by N. Balaji

Article E-Mail :: Comment :: Syndication

Stories in this Section
Big guys do it better

Make way for more space

Doing away with daemons

Access denial

Distance makes heart fonder, data safer

Technology isn't the hurdle

No walls in the IS world

Quiz

A dash of Zen for more profits

Cartoon