The Hindu Business Line : Bounce back

A computer glitch in a global marketplace this month highlights the vulnerability of large data systems — even as system and storage security players join to even the odds of recurrence..

Anand Parthasarathy

A catastrophic computer crash on September 9 brought trading on the London Stock Exchange to a halt for seven hours. It prevented thousands of online traders from doing business on what they hoped would be one of the most profitable days of the year: as news of the US government’s bailouts of the mortgage giants Fannie Mae and Freddie Mac triggered heavy global trading. Angry brokers and financial houses fumed as other exchanges from Tokyo to Frankfurt to New York scooped up the lost business.

“When ‘latency’ — the time it takes to execute a trade — is measured in milliseconds, seven hours might as well be seven light-years,” wrote the Financial Times acidly. Why did the world’s second-largest exchange not have a successful back-up system (or if it did why did it fail), asked a broker.

Experts called it a ‘black swan’ event — that is something that could never be foreseen — and it had reportedly to do with connectivity.

At around the time when London’s financial world woke up to its ‘outage outrage’, experts from the world’s leading computer storage and system management companies were congregating in Singapore at the annual Vision Summit of system security leader Symantec. On parade in the showcase that accompanied the conference were dozens of solutions that touted buzz words such as ‘end point protection’ — data security at the devices professionals deploy to access their corporate networks, such as laptops or mobile phones — or ‘data recovery’ — ensuring that large systems bounce back to business in the shortest possible time after a ‘data disaster’.

The eight-hour time difference — ahead of London — ensured that news of the computer crash reached the instant update services of the Blackberry-wielding executives in Singapore only as the conference was in its last stage.

Otherwise, the assembled software professionals and customers might well have asked why the highly promoted solutions and scenarios to ensure data continuity did not work even when there was no catastrophic breakdown of power or any natural disaster. Some of the companies known to have helped create the London Stock Exchange electronic trading system and its innovative electronic order book solution, both said to be the fastest in the world, were among the key players at the Symantec event and it would have been embarrassing if they were to have been interrupted in the full flow of their eloquent presentations of fool-proof data systems.

But the incident, albeit in a perverse way, highlighted what the theme of the Singapore conclave was all about. And Dr Basant Rajan, Chief Technology Officer, India, of Symantec, put it bluntly: Data damage could occur on an ordinary day, in a mundane way — even without headline-hogging natural or man-made disasters. It was — sadly — inherent to all automated operations and “failing to plan how to tackle such events was like ‘planning to fail’,” he said.

He had the numbers to back his assertions: he had overseen portions of Symantec’s latest IT Disaster Recovery Survey, the results of which he presented. The sobering fact was at least one in three Asian corporates polled for the survey was ill-equipped to handle data disaster if it occurred. Another one-third actually had to deal with such episodes — but sadly there has been a significant decline in higher level executive involvement in disaster recovery plans — they tend to be relegated to ‘lesser mortals’ in the organisation.

What caused the episodes in the first place? The study said: Hardware and software failure (36 per cent of organisations); external security threats (28 per cent ); power outage/failure/issues (26 per cent); natural disasters (23 per cent); IT problem management (23 per cent); data leakage or loss (22 per cent ); and accidental or malicious employee behaviour (21 per cent of organisations).

Progress, its own enemy sometimes

One revealing fact of the survey showed that sometimes progress can be its own enemy: Virtualisation, where the resources of multiple storage or server units are pooled and viewed as a single logical unit, is the latest panacea seen as bringing significant efficiencies to enterprise systems. But implementing virtualisation is being cited as a major reason why over a third of corporates interviewed are having to re-evaluate their disaster recovery systems. In other words, the super remedy brings with it a minor hurdle that has to be factored in.

And while data loss could have a thousand causes, data recovery, especially after a disaster-level episode, finally boils down to doing every thing twice — as simple as that.

Back up all critical data, then store it in a smart way, some of it in a different physical location to overcome a natural disaster; some of it close by in another media, near enough for fast restoration of the system if the main systems pack up. It is both art and science — and there were many solutions on offer to help enterprises create a fail-safe system or get very close to one.

Ways of going about it

Hewlett Packard has relieved customers from the ‘dharma sankat’ — tape or disk for back-up — by letting them have their cake and eat it too. It offers a D2D2T system, that is disk-to-disk-to-tape, in its StorageWorks line-up that offers users two tiers of back-up with differing media. It has also taken Symantec’s portfolio of antivirus and anti spam tools and built them into a software service where one can funnel all incoming e-mail to HP, which will serve as a filter before it reaches one’s own servers.

The King of the Ring in virtualisation is VMware — but in addition to its own large portfolio of Infrastructure Automation solutions, it also joins players such as NetApp who offer back-up solutions for VMware environments.

Backing all data places excessive load on the computer’s processing power. In a virtual environment one can achieve significant economies by what is known as Data De-duplication. Simply put, this means: if I create a great power point presentation and mail it to 20 of my associates in the company, should the PPT be saved 21 times on the server that caches all corporate mail — or will some clever software detect the duplication — and store just one copy? An acknowledged leader in the de-duplication business is Data Domain — and the Santa Clara-based company recently launched its India operation, with a Bangalore base.

Simultaneously it has brought its latest flagship product DD690 — a scalable de-duplication storage product to Indian data centres. No confusion at Data Domain about tape versus disk. It has hitched its wagon firmly to the disk trail.

IBM’s predictive failure analysis and diagnostic systems sniff out a potential disaster before it can happen; Sun’s and Tandberg Data’s storage solutions remain agnostic in the disk-tape debate, with offerings in both media; Fujitsu straddles the server-storage world with parallel offerings: Primergy virtualised servers and Eternus scalable storage offerings... the ecosystem grows.

“Every one has data and any one can lose it!,” says Michael Wolfe. He is the founder and key architect of Vontu, a firm that offered for the first time a solution dedicated to Loss Prevention — an integrated suite to prevent loss of confidential data wherever it is stored or used, across end devices, networks or storage systems. You have company confidential information on your laptop? You have just lost it? How do you ensure the data can never be accessed by any one other than yourself?

Vontu provided the comfort of knowing your data was safe — even after you thought you lost it. Six months ago, Vontu became a part of Symantec and today Vontu DLP 8 (for Data Loss Prevention) is an integrated offering that seamlessly works with other endpoint protection tools.

T.G. Dhandapani

DLP of the Vontu kind is part of the continuing technological challenge — how to keep your systems up and running for days, months, years — and if disaster strikes, how to get up and go again in milliseconds rather than hours.

If data is misplaced, how to ‘lock’ it ‘For Your Eyes Only’. This is a business where only the naive say it won’t happen to them, where the wiser ones say ‘Never Say Never Again’. Those enterprises whose heads are old enough to have read a steamy novel of the 1970s by Jacqueline Sussan will do worse than adopt her title as their motto: ‘Once is Not Enough’. Save all your critical data twice.

Symantec awards

Pravir Vohra

Two Indian enterprises were honoured at the recent Symantec Vision Summit for their innovative use of security products and services.

T.G. Dhandapani, Corporate Chief Information Officer at the TVS family companies, TVS Motor Co and Sundaram Clayton, received the award for the successful and effective deployment of Symantec End Point solutions across the entire corporate and dealer network. The system covered 3,000 desktop machines, 500 work stations and 100 servers, Dhandapani told eWorld. The system is fuelled by SAP.

Pravir Vohra, Group CTO, received the citation on behalf of ICICI Bank for implementing Symantec security management solutions that are expected to provide the bank $2,50,000 in savings over previous solutions, as well as 10,000 hours of manual work in tackling spam.

Related Stories:
Under siege in cyberspace
Serious about security