Data management –
key to RFID’s future success
The market for Radio Frequency Identification (RFID) technology is beginning to take off. It is projected to grow by 255 per cent from $900 million in 2003 to $2.3 billion in 2010, according to Future Horizons, a UK-based market research firm.
RFID as a technology is not new, however, advancements mean that this technology can now be applied much more widely with real business applications. In the last twelve months numerous organisations have embarked upon trials with positive results, this now means that organisations will start applying this technology where they see a business advantage.
The knock-on effect of this is that new applications will have to be built to serve these business functions, and existing applications re-architected. There will be a number of challenges that will have to be overcome, for example, an RFID tag typically identifies an individual product item or transit container, and therefore RFID systems will generate much larger amounts of data than exists within barcode-based systems.
InterSystems offers a rapid application development and high-performance database platform that is capable of handling the vast amounts of data which RFID applications will generate. Additionally, InterSystems offers a services oriented and event driven, integration platform solution which is key to integrating RFID information with the rest of the enterprise.
The RFID market is such that no one company can provide a complete end-to-end solution, this means that developers will be key to any organisation that is considering introducing RFID applications into their business. In addition, these developers will need to work more closely with other technical and operations staff.
RFID projects need to focus on the real cost of business benefits – the greatest benefits may require changes to business processes. These changes will be determined as the pilot phase develops, this means the development process and application will need to be flexible enough to cope with the changes that will take place throughout the pilot and leading up to production.
Unlike typical business applications, RFID applications receive physical measurements – tag reads – in (near) real-time; these are known as events. This is where the large volumes of data will be created within RFID applications. Business events will also be generated, which will need to be communicated to existing systems, ideally in real-time.
To provide useful business information, the raw tag events need to be filtered to ensure that business decisions are based upon reliable and accurate data. One way to achieve this and relieve the volume of data hitting these applications is to develop an event filter to select only important events, such as firm tag arrivals and departures. Filtering requires access to data, such as product information, as this data is not held on the tag; this is provided by information services which are typically external systems.
RFID applications are likely to be Composite Applications as they combine new technology (RFID) with existing working software. These composite applications will involve a mixture of packaged applications, hardware and existing applications which the developer will need to integrate, while building new functionality.
Managing the data explosion
RFID applications are able to capture vast amounts of event data that has not been available to the business previously, and so the business will now need to determine what information it needs, to what detail and when they will need it.
This can therefore potentially result in huge pressure being put upon RFID applications in terms of the amount of data they need to capture, filter and store.
Two options available are:
• Use an in-memory database and an external data repository, with filtering which must reduce the amount of data to a level which the external repository can manage.
• Choose a highly scalable data repository, with filtering more focussed on business needs.
Besides the event data there are two additional types of data used in RFID applications; these are:
• Static and slowly changing data – e.g. product information
• Dynamic data – e.g. data for tracking items through the supply chain
Where product information is sourced from external systems, it may be necessary to cache this in a local high performance repository to meet the real-time performance requirements.
The InterSystems solution
The RFID requirements can be broken down into three main categories:
• The need for rapid and flexible application development.
• High powered data repositories.
• Event Driven and Service Oriented Integration.
InterSystems has two main product offerings that fulfil all of the above challenges, namely Caché and Ensemble.
Caché is a high performance database management system with the performance of an in-memory DBMS. In addition, Caché provides rapid application development through its Unified Data Architecture that allows the developer to build applications using both objects and SQL technology.
Caché’s integrated development environment supports this simultaneous access to data using both object and SQL technology, thereby supporting rapid development without sacrificing performance, and providing the flexibility needed for RFID development. Persistent objects can be used to access local RFID static data, dynamic data, and persistent event data directly. Transient objects may be used for business logic, event filtering, and transient event data. Caché supports fast inserts and updates with simultaneous fast queries, to support real-time event capture and processing, which has been proven in the transaction intensive finance and telecommunications sectors.
Caché supports a variety of development technologies including Java, COM, .NET as well as an in-database object scripting language; since it also runs on a wide variety of platforms, the developer has the freedom to choose the appropriate technology for his/her application.
Caché provides very low latency for database operations which is necessary to meet real-time performance requirements so that the developer does not have to code around database limitations. Its scalability technology can be used to implement highly scalable information services. Caché is built for high availability and Caché systems can be developed and managed with minimal DBA involvement, meaning that systems can be built and deployed rapidly.
Ensemble is a comprehensive and architecturally consistent Universal Integration Platform that enables fast integration and rapid composite application development. Its Unified Service Architecture allows new business solutions to integrate data and orchestrate processes while retaining the value of existing applications.
Ensemble is a single integrated product – not a suite, making it is easy to learn, use and maintain. The integrated development environment provides editable business process diagrams, which allows the developer to quickly build business process applications and use RFID technology to automate previously manual processes. Such applications are inherently flexible, as the business process components are easily reconfigured. This business process technology is ideal for managing the integration with other technologies, third party software and existing systems.
It is important to realise that some external systems may not be sufficiently responsive to support real-time interactions; Ensemble provides queued asynchronous messaging for talking to these systems without impacting the core application; the messages are coordinated by business processes which are orchestrated by Ensemble. A single business process instance can be waiting on multiple external systems simultaneously. In this way an Ensemble application can act as an intelligent buffer between RFID readers and external systems.
With its built-in high performance data store, Ensemble persists all messages and this has many benefits such as reliable delivery and message tracing.
Its low-level queuing and event notification technology can be used to implement a high performance event filter for raw data events; the filter can generate application events as Ensemble messages, which are further processed by business process components before transmission to external systems. Raw events can also be stored directly in its persistent object engine, for diagnostic or other purposes.
Both Caché and Ensemble provide built-in support for XML and Web Services for integration with other systems using them. The XML technology is integrated with Caché Objects, giving three-way integration between XML, Objects and SQL, which makes life easier for the developer.
Consider the potential business value of all those rich messages stored in Ensemble. As an added bonus, SQL queries may be used to mine those messages to generate new application events which can be fed into business processes along with RFID event data. This provides a “feedback loop”, which allows the business value of the messages to be realised. This is just one aspect of the Business Activity Monitoring technology provided by Ensemble.
InterSystems has been working on a number of RFID projects, in particular we are engaged with a large multi-national logistics organisation in conjunction with their end user retail clients. In trials we successfully demonstrated:
• Real-time capture and storage of all RFID events.
• Event filtering network based on Savant architecture.
• External integration managed by business process components.
• COM integration.
• Real-time display via web interface.
• Very fast queries against stored raw event data.
• Correlation of transit container events with vehicle events to determine vehicle direction and contents.
It is expected that production systems based on the trials will be deployed at retail clients early in the New Year.
By Andrew McLellan, sales manager, InterSystems.