By enabling objects to inexpensively transmit information about their location, configuration, and history to a remote automated reader, RFID technology promises to radically improve the efficiency and effectiveness of various business processes from warehouse inventory control to security. Furthermore, RFID chip manufacturing costs have dropped during the past year to the point where widespread RFID deployment has moved into the realm of economic feasibility. Nonetheless, only 10% of Global 2000 organizations are currently engaged in RFID pilot projects. We believe approximately half of these projects are already being written off as failures, while most of the rest have yielded indifferent results. Next year will likely see a drop in the level of RFID investment, while many organizations pause to re-evaluate the technology's true potential.
Short-term slump but long-term promise
Although the current dampening of expectations surrounding RFID is a natural reaction to the almost hysterical enthusiasm that characterized discussions of this technology during the first half of 2003, organizations should still prepare for RFID's pervasiveness in the longer term. By 2008, we anticipate that approximately 30% of manufactured capital goods will be RFID-enabled with that percentage growing to 80% by 2013. RFID chips will also find their way onto consumer goods, though privacy issues will ensure that progress in this sector is slower.
IT organizations will be particularly burdened by this technology's widespread deployment because the monitoring, filtering, analysis, and much of the first-line response to RFID-generated event signals will fall on their shoulders. At the same time, the explosion in data volumes and the ensuing need for monitoring and management will fundamentally transform the infrastructure and application management market, as vendors race to extend their respective product portfolios' sense-and-respond functionalities into the RFID space. Although both Computer Associates and IBM Tivoli have had some experience monitoring and managing non-computational objects, both vendors' technologies depend on minor extensions of an SNMP-based classical manager/agent architecture. However, the structure, volume, and frequency of RFID signals are sufficiently different from anything resembling an SNMP environment that any design lessons such vendors learned from past efforts are unlikely to be applicable in an RFID setting.
The three obstacles
Before RFID deployment resumes its growth, several obstacles need to be overcome. First, the market will require standardization and stabilization, with regard to both the actual technology and functionality deployed within the various RFID chips and the syntax and semantics of the RFID signals themselves. RFID technology evolution has accelerated only recently, with vendors shifting from a focus on simple repetitive signals transmitted over short distances (and evaluated primarily within a bar-code replacement context) to a focus on adding intelligence to objects (i.e., allowing for a more complex signal grammar, communication over longer distances, and machine-to-machine conversations). Even more problematic, trade conflicts between the European Union (EU) Commission and the US are preventing stabilization with regard to the syntax and the semantics of the most basic RFID signals (e.g., the differences of approach between the European Telecommunications Standards Institute and ANSI). At present, global organizations face the prospect of having to monitor, manage, and translate between two different signal grammars depending on how their intelligent objects are distributed among the regions of Europe and the rest of the world. Although the slackening of interest in RFID will restabilize feature/function sets, inter-regional trade friction in this area is unlikely to lessen for at least five years, adding significantly to the long-term costs of RFID deployment and management for global companies.
Second, RFID technology's potential ability to enable automated gathering of information about post-sale product usage has already stirred up privacy anxieties. Legislation limiting or even preventing the deployment of RFID has begun to wend its way through both national and regional legislative bodies. Although we ultimately believe perceived economic and security-related benefits of RFID deployment will trump privacy concerns, usage protocols that minimize the scope for intrusive consumer observation will need to be developed, and this will slow deployment.
Third, RFID signal and reader technology - taking into account some of the more recent enhancements previously mentioned - operates at a low semantic level. At the same time, the volume of signals generated is large. To actually make use of much of the information contained within the signals, vendors will need to develop (and organizations will need to implement) a rich two-dimensional infrastructure. The first dimension will be devoted to management of the signal traffic - overseeing the entry and departure of objects from signaling environments, ensuring flow, encrypting transmissions, minimizing collisions, and protecting corporate networks from unanticipated surges in RFID-related traffic. The second dimension will be devoted to the elicitation of meaning from the signals themselves and providing the linkage between RFID signal generation and other corporate applications. In other words, RFID is not a self-contained technology. Although systems integrators could create the required middleware (and we anticipates that, during the next three years, RFID-related projects will generate approximately $2B for the systems integration market), widespread deployment will presuppose off-the-shelf availability of RFID management and interpretive software, an availability unlikely before 2006.
As the previously mentioned time frames imply, the third obstacle will be the first to be overcome. In fact, we anticipate that approximately 20% of new IT and communications-related venture capital investments in 2004 will be in the RFID space. At the same time, despite the current furor around privacy, the modifications of RFID technology required to ameliorate concerns are minor. The most enduring obstacle will be conflicting syntax and semantics standards now being contemplated by the US and the EU. Although we believe the long-term benefits of RFID are such to make costs associated with dual standard support worthwhile, failure to come to a global agreement ensures that deployment will be limited until the end of this decade.
Bottom line: The short-term prospects for RFID have been overhyped. Serious obstacles stand in the way of widespread deployment. Technology will become a significant factor in various industries in three to five years.
Business impact: Organizations must avoid short-term investments in RFID infrastructures. They should wait for standardization, a resolution of current political/legal conflicts, and the emergence of off-the-shelf RFID middleware.
Click here for more business and IT research from META Group
Copyright ©2004 META Group Inc.