Moving towards a linked data environment

Linked-Data

A wealth of valuable data is increasingly being stored on internal networks in major organisations and enterprises across the UK without its full potential being realised. In the current economic climate, as budgets are being rapidly cut and competitive pressures rise, it is time for organisations to unlock the inherent value in the data that they already own. By doing this, organisations are able to drive business efficiencies, and improve quality of service.

Why do we want to link data?

All data is some partial representation of some aspect of the real world. CRM systems have data models of people and customers, whilst support departments have models of customers and support cases. Accounting has models of customers and which products they buy, whilst engineering has data about products and related documentation.

All data of this type usually resides in custom built or off-the-shelf systems that both manage the data and provide specific functionality. For example, in the case of accounting, CRM systems provide accountants with the data required to invoice a customer. While this supports the local requirements of a given department, the data is not available for other business functions.

Now imagine a world in which all of this data was available as one unified data model. For example, a person in accounting would be equipped with the tools to find out if the person they were about to call about an unpaid software subscription, had had lots of open support cases. Further to this, a sales person using linked data would be able to find out the exact numbers of sales of a product to a given market segment, based on invoice data.

What does it mean to link data in an enterprise?

Linking enterprise data is not about pouring it all into a single new data repository. Linking data is about exposing the raw data from each of the different systems using RDF and HTTP. Every business entity, person, product, customer, location, service, is given a unique URL that when resolved over HTTP, returns the relevant data for that entity in RDF. Each silo that exposes linked data would also typically expose a query endpoint.

The linking part of this comes from the fact that information in one data set points to or links to a data entity in another linked data wrapped silo. For example, when pulling back the data for a customer from accounting, it will contain order information. The products purchased will be listed as links to the product management system. This system can return data about the given products and may in turn link to the human resources data for the lead engineer.

Don’t we use SOAP for integration?

SOAP web services have been used in the past to provide application integration across the corporate network. The main drawback of this approach is that as requirements change new SOAP RPC calls need to be implemented. This is costly and slow and the business advantage of having the data may have passed before the feature is implemented. With linked data, the premise is to make all data available from the beginning, as it is hard to predict how other departments can make use of it. This allows other users to derive value from the data that wasn’t conceived of by its owners.

Empowering people

Linking data has some key benefits, of which the most valuable is the ability to deliver to employees a holistic viewpoint of the data entities that they deal with every day. Employees are better informed to make decisions when they have contextual information at their fingertips, and by linking data, it is possible to reduce the time it takes to find this relevant information.

Computer processing

As well as enabling people to better perform, it also allows machines to process data as a whole. Reasoning and processing can be carried out with an internal linked data system, where before it would have required a large and dedicated project to complete the task. If a company can make its data available and linked, it would be possible to run a process that audited when the support calls in a given market segment was very high. Further to this, it could highlight this to the sales team.

Call to action

In today’s economic climate, businesses need to make use of the data and information they have already. By creating a linked data environment, they can connect and expose the existing enterprise data in way that can be used by both people and machines. Businesses would be able to make more informed decisions and carry out analytics that weren’t possible before.

Graham Moore is director at NetworkedPlanet. He's a passionate and driven professional who seeks to deliver high quality technical solutions in the area of information and knowledge management.