Just 20 years ago, everything that meant anything to an organisation would have been in a database, says Phil Tee. In five years time, he says, less than 10 per cent could be there – the rest will be files on employees’ hard disks or corporate storage systems.
This “unstructured” data is a growing problem for big organisations and by some estimates is expanding at 70 per cent a year – a reflection of the distributed, interconnected business environment of the 21st century.
Mr Tee is co-founder, chairman and chief technical officer of Njini, a small UK company that has developed what it claims is a unique solution to the problems created by unstructured data.
“Structured data [as found in a database] has management built in,” says Mr Tee, who is best known as the co-founder of Micromuse and founder of RiverSoft. “Unstructured data has chaos built in. What we are trying to do at Njini is place the management there, in a way which doesn’t force the user to change their habits too much.” Last June, after two years of development work, the London-based company launched NjiniEngine. This is claimed to be the first product that identifies and tags data intelligently and at the point of origin, working automatically and in real time. The software can determine where the data should be stored, for how long, who should be able to access it and how many copies should be kept.
The software uses the metadata (data about data) that is generated with every file, such as information an Office document might carry about whether it contains pictures or whether it is marked “commercially confidential”. It then attaches business policy to it, wrapping it in a “jacket” that stays with it throughout its life. All this happens before the document is stored.
So, for example, the policy could dictate that nobody outside the legal department can access the document, or that it can be printed only on a certain printer to restrict hard copy distribution. Alternatively, proliferation of non-business data such as holiday snaps could be enforced with a policy restricting pictures using the jpeg format to the graphics department.
Mr Tee says all the “head-scratching” encapsulated in Njini’s patents has gone into ensuring that the software works in a way that is scalable, and to ensure the user does not notice any degradation in performance.
Classifying unstructured data according to business policy is a hot topic in the enterprise IT sector, with other small companies such as Kazeon, Arkivio, Scentric and StoredIQ all trying to establish positions.
Mr Tee says Njini’s differentiator is its ability to work in real time. A lot of policy management systems run in batches, perhaps at night some hours after the files have been created, he says. “This creates an opening in the system, and if you are trying to enforce something it’s a non-starter.”
The reception for NjiniEngine suggests the company is on to something. Last month the software won the 2005 Product of the Year prize at the UK Information Management awards.
“Njini looks to have built an elegant answer to the question of information asset management,” says Jason Stamper, editor of Computer Business Review and a judge at the awards. “For all sorts of reasons, not least the latest compliance requirements, companies need to start treating business data just like any other asset that needs to be managed throughout its life cycle.”
David Jones, a Silicon Valley veteran who joined Njini as chief executive in November, says it is focusing initially on NjiniEngine’s ability to prevent duplication of files and thus reduce corporations’ storage costs. “The value proposition is drop-dead simple – payback can be measured in months,” he says.
Beyond that, by helping identify the business value of data as soon as it is created, the software could bring benefits wherever “white-collar productivity is a key part of the production process”, says Mr Jones. This could be any big organisation that produces lots of documents, although initially Njini is targeting the financial services, healthcare/pharmaceutical and government markets.
Customer: Virgin Mobile
Among the early customers for Njini is Virgin Mobile, which is “pretty close” to a full commercial purchase of NjiniEngine after conducting a successful proof of concept, according to Keith Bennett, the mobile operator’s lead infrastructure architect.
Njini, which had been in contact with Virgin Mobile about some earlier product ideas, approached the company again last spring. “We really liked the idea [of NjiniEngine], and wanted to be involved,” says Mr Bennett. “We have been helping them steer the product slightly in a few ways, especially on the operational aspects.”
Mr Bennett says the software allows Virgin Mobile to achieve, in one box, everything it wants in terms of reducing duplication of files and managing unstructured data.
“The key thing, first and foremost, is that the user is completely unaware of it,” he says. “But if user A and user B save the same file into their shared drive, it works it out and stores just one of them, immediately saving space.”
The software could also be used, he says, to set a policy on how long jpegs and images should be retained in the first tier of storage, before being moved automatically elsewhere. As a brand-aware Virgin company, the operator generates a huge amount of images, but often they are never touched subsequently.
By cutting down duplication and setting policies like this, Virgin Mobile could reduce the need for extra storage and may not even have to buy any this year, says Mr Bennett.
“We are not a massive organisation, and we don’t have tons of storage, but we have identified areas where we can reduce costs and this product will give us savings,” he says. He believes the software could pay for itself in about a year, and definitely less than 18 months.