Talis continues semantic evolution with Kasabi
'Data marketplace' from Birmingham's semantic web pioneer offers cloud-like functionality, but will it bring linked data to the masses?
In 1969, a group of libraries in Birmingham decided they needed to become more efficient. Calling themselves the Birmingham Libraries Cooperative Mechanisation Project (BLCMP), the group built a centralised database of ‘machine-readable’ bibliographic records, first using microfilm to store book data then, from 1982, using IBM mainframes with terminals at each library.
BLCMP went on to become Talis, named after its integrated library system, and for many years it was leader in the automated library management software market. But that is a mature market, and last year Talis divested its library division to focus on the company’s other passion: semantic technology.
Since 2007, Talis had been developing a data hosting platform that allows the customer to describe the meaning of their data using the resource description framework (RDF) standard. This standard allows each item of data to be identified by its meaning – whether it is the name of person, for example, or the title of a book.
Two significant customers that have published data on the Talis platform are the BBC and the government's open data portal, data.gov.uk.
Both those accounts are high-touch, consultative engagements for Talis, but in 2011 the company identified demand for a more cloud-like, self-service platform through which to access the technology. In March of that year, it launched Kasabi.
“The idea behind Kasabi is to have a self-service environment where anyone can sign up, and get a hosted data store with a range of [application programming interfaces] APIs,” explains Leigh Dodds, Kasabi’s CTO.
The service allows organisations to publish their own semantically tagged information as ‘linked data’ – i.e. each item has its own permanent link – so that developers and other users can access it over the web.
Kasabi can be used on a pay-as-you-go basis or there is an enterprise license, which costs £1,500 a month and allows the customer to publish 500 datasets and make 5 million requests for data.
The original ambition – and still the ultimate goal – was to create a unified, web-based data marketplace that would allow developers to combine the data, build new applications and spot new correlations.
However, Dodds and his team have found that some organisations are more interested in using the technology powering the service than the data marketplace itself. Kasabi is therefore now licensing its infrastructure to customers that wish to create their own web-facing semantic data stores.
“We’re providing Kasabi as a white label option, so companies can start to open data up under their own brand and their own domain,” Dodds explains. “Then we can support them exploring value generation around that, whether its direct monetisation of the data, or driving traffic to other products or services. We’ve found that we needed to take a step backwards and help people over that final hurdle of opening up their data,” he says.
Dodds declines to reveal how many customers are using the Kasabi service. It is very early days, he says, and most organisations are still wrestling with the concept of opening up their data.
However, the cause got a boost in December 2011 in the form of a report on linked data from analyst company Gartner – a potential seed of mainstream interest.
Presenting both the BBC and data.gov.uk as examples, the report described linked data as "the best way of opening, integrating and sharing data to meet the needs of the evolving web".
“Business benefits offered by the use of linked data include revenue generation, deeper connections with customers and easier compliance to transparency mandates,” wrote author David Newman. “Technology benefits include bridging silos, delivering a web- oriented method for data integration and uncovering previously inaccessible data sources.”
Newman predicted that linked data will reach mainstream adoption in five to seven years.
Being five to seven years ahead of the adoption curve is not a bad achievement for a 43-year-old company. However, Talis and Kasabi may need to undergo a few more evolutionary steps before their true potential is realised.