The Laboratory Information Management Systems (LIMS) sector is a mature one, but nevertheless seems poised to make respectable growth over the next few years. Market research published by Frost and Sullivan and ARC Advisory Group forecast that the worldwide market could expand at between five and seven per cent per annum. The growth is coming from life science, medical, and other firms that are trying to break down the barriers between separate parts of their businesses as they seek to become more efficient. This may be because of management efforts to globalise a business, increased competitive pressure, or - in the pharmaceutical industry - increased scrutiny by regulators such as the US Food and Drug Administration.
These companies increasingly want software systems that provide economies of scale across their business, increasing productivity and improving product quality. There is greater need to integrate LIMS and other specialised software with enterprise resource planning (ERP) systems such as SAP, says Paula Hollywood of management consultants ARC and author of a report on the future of LIMS. 'In order to be effective, critical information subsets maintained in ERP and LIMS records must be shared and synchronised, such as sample testing, specification targets and limits, revision numbers, and dates,' she says.
In the past, companies have sometimes shied away from upgrading or replacing software because they believed that commercial LIMS systems needed extensive customisation. That view is outdated and, now, says Hollywood: 'Manufacturers are displaying an increased willingness to implement moderately customised, out-of-the-box solutions to decrease their LIMS Total Cost of Ownership (TCO).'
By adopting open web-based standards, LIMS manufacturers have developed products that are more easily customised and offer much more than sample management and reporting functions. These standards should allow systems to link to wider business processes, to share scientific data throughout a worldwide organisation through some central hub, but without the extensive customisation of old. To achieve this, LIMS vendors have seized on the latest incarnation of distributed computing - web services - which is heir to middleware platforms such as CORBA (Common Object Request Broker Architecture) and DCOM (Distributed Component Object Model).
'Web services give us a better opportunity for integrating with the environment as a whole,' says Adrian Fergus of Thermo Electron's informatics division. 'They allow you to create a generic interface to multiple systems, rather than a point solution'. This makes it a lot easier to share data around a company than the application-to-application communication possible with traditional middleware. And they have been adopted by all of the major software players, says Fergus.
Put simply, web services are software applications that can be called over a network. They are modular and can be located, or invoked from, anywhere on the network and are built using XML together with a group of messaging and other standards (both old and new). URLs give the address of a web service and HTTP is used for communication between applications. The newer standards have been developed by a number of companies working to create a set of platform-independent protocols based on XML. The web services description language (WSDL) describes what an individual web service does: that is, all the methods available to a client application. The Simple Object Access Protocol (SOAP) is a standard for XML-based information exchange between distributed applications, while the Universal Discovery, Description and Integration (UDDI) registry is a sort of directory storing details about companies and their web services.
Companies can then, for example, run web services on a server based around one operating system (e.g. Linux or Windows) and call them from another software application running on a different platform. The application can be written in any one of a number of programming languages (e.g. Java or Perl) and can be a browser-based client or one with its own interface. All that is needed is the ability to work with XML and to communicate with an internet URL or an internal intranet server.
An architecture for web development
Web services began to emerge in 1999 as various groups moved to explore ways to combine HTTP and XML and to provide for ways to exchange XML documents. Starting as independent projects, the World Wide Web Consortium (W3C) moved to bring web services under its banner in 2001 and to set out an architecture for their development. Much ground was covered in 2002 filling out that vision and preparing web services for the mainstream. In February 2002, IBM, Microsoft, Intel and others formed the Web Services Interoperability Organization (WS-I) to promote standards, and to do this not by establishing new protocols but by assembling profiles - sets of standards that work together - from the W3C, OASIS (Organization for the Advancement of Structured Information Standards) and other bodies. Of course there has been the usual jockeying for position - between standards-setting bodies such as OASIS and W3C and the larger companies like Oracle, Sun, IBM and Microsoft. But the standards have stayed open and gained enough supporters to put them in the mainstream.
Computer scientists are still arguing about the relative merits of web services over CORBA and DCOM but web services can sit beside these older - and costly - approaches to distributed computing. And nobody can deny their increasing popularity, much of which is based on the potential for improved interoperability. A report published by IT market research firm IDC in June forecasts worldwide spending on software in support of web services-based projects will reach $11 billion by 2008, compared to $1.1 billion in 2003. 'The industry is entering a critical stage in the acceptance and support for higher-levels of web services standards and technologies. The market-at-large is embracing the concept of services-oriented architecture (SOA) to such a degree that it could inevitably leave a mark on IT environments for many years to come,' said Sandra Rogers, director for web services software and integration at IDC.
The overall driver for the take-up of XML-based Web services is greater software and data integration, but the aim isn't simply to provide more widespread access to data by adding web interfaces to existing software. What companies are really looking for are increases in productivity through greater efficiency, by tighter integration of the systems used throughout an organisation - that is, between LIMS and other business applications. Companies hope also to standardise software, or at least protocols, across sites in many locations. In part this is driven by the need to increase productivity, but a major reason is that mergers and acquisitions have left many firms with a motley collection of software to maintain. Colin Thurston of Thermo Electron sees this as the reason for 'the biggest shake up in the LIMS market right now.' Companies running big international research programmes 'see the tools that facilitate those global projects on a global basis,' he says.
Increased competition means that companies don't want to spend too much money achieving these goals. In fact many of them want to make significant cuts in their IT spending, despite improved economic prospects, at the same time as seeking to improve software interoperability. In their sights is the 30 per cent of their budget (a widely used figure, cited by Thermo Electron) that is spent on system integration. Web services enthusiasts say the technology could square this circle by making things simpler and ultimately cheaper. When all the applications used within an organisation employ a common standard, such as Web services, it is easier to include new applications or replace or upgrade old ones. 'If you upgrade one end of the interface you're not going to affect the way that it operates. You're not trapped into having to upgrade the middleware or the other [client] end,' says Thermo Electron's Fergus. The architecture of web services and its open standards thus spares the cost of major integration work with complex middleware. The effects in a large international organisation could be dramatic, with data being exchanged quickly and easily between locations.
LIMS companies have already begun including web services modules in their software. Starlims has established its Web Services Framework, which makes available key LIMS elements such as sample tracking, results entry, and a variety of quality-related information for distribution to other systems. These services are made available to enterprise-software, running procurement, logistics, human resources, regulatory reporting, or other processes.
Thermo Electron initially offered web services solutions around two years ago, but as a service rather than a product. In the past year, the company has released products - Nautilus and Sample Manager - that support these services natively. In Sample Manager, web services are even used to integrate modules of the software together. Web service-enabled Sample Manager 2004 was released in February and given its first public outing at the Pittcon laboratory equipment exhibition in Chicago in March. The software also boasts a new instrument-integration capability based on Microsoft .NET technology.
LIMS and ERP get closer
With Sample Manager, Thermo Electron's target is large-scale production environments, where Enterprise Resource Planning (ERP) systems such as SAP Software's R/3 Enterprise are the standard hub around which other systems are grouped. LIMS and ERP systems have been growing ever closer together - for example, SAP promotes its product lifecycle management (PLM) solution as an alternative to a traditional LIMS - as both ERP and LIMS users have sought access to information from each other's systems that is immediately usable, with web services facilitating the two-way transfer of information.
Web services are still under development, and there are unresolved matters relating to some standards. 'It is important that a standard schema is adopted for XML for handling quality data and analytical data,' says Thermo Electron's Colin Thurston. 'There are different ways in which some science-based industries are using XML. It is affecting the level of integration we can get to'.
The development of appropriate XML standards could expand the potential of web services to integrate scientific software, particularly in the R&D sector. 'Chemical information management systems, which handle chemical structures, need to be interfaced to LIMS', says Thurston. These already have a standardised format, based around XML. More challenging would be the creation of a hybrid electronic laboratory notebook (ELN) that interfaces to a LIMS, analytical data systems, a document manager, authoring and archiving software. The result he says: 'You don't have a single ELN product, it will be an integration, with web services acting as the glue to hold all these together.'
