ObjectWeb RFID Workshop, Middleware Conference
There is an ObjectWeb workshop today, at the Middleware 2005 Conference, about FRID. David Li (board member of OW representing individual members) presents a case for an ObjectWeb RFID Initiative.
RFID (Radio Frequency IDentification) are small tags able to transmit unique ID through radio. They are expected to replace Bar Codes in the 10 years to come. Mandates from Wal-Mart, DOD, Target, Metro is a driver for adoption of RFID and next wave of IT business. A RFID can be attached to virtually anything, including human beings. Nokia recently introduced the RFID scanner cell phone. An application may be buying goods in vending machines while having their price imputed directly on your phone bill.
A vision of the future of RFID is an "Internet of Things"; tags everywhere, readers everywhere. It would be necessary to leverage the Internet to carry data. Current RFID tags (namely the EPC class of tags) have limited capacity of 96/128 bits to store their (unique) id. Providing a network infrastructure falls into the scope of ObjectWeb.
Exemples of applications include tags on boxes, tags in passports, fight against counterfeit drugs, livestock, pets, kids... Sounds to me that nobody wonders whether this future would be a Brave New World or just... hell.
Gartner's Hype Curve analysis puts RFID Peak of Inflated Expectations in 2004, Trough of Desillusionment in 2006 and Plateau of Productivity in 2018. Market size in 2008 is expected to be around $6.5 billion. The big players are the usual suspects: IBM, Microsoft, Sun, Oracle, BEA and RadioActive & FirstOpen in open source. In the Java space, two JSRs have been proposed by Nokia about the reader interface.
In the architecture, data filter and transport is an important aspect, because unlike bar codes, RFIDs are constantly on and the readers keep reading information. This is somewhat similar to what happens with mobile (cell) phones. Management of RFID infrastructure extends beyond managing the reader. For instance, it would also include management of sensors to detect moting of packs/trucks transporting the goods. The information network is necessary to share information in and between organizations.
EPCGlobal is leading the current specifications. It's mainly a hardware specification, American centric (generation 1 uses 900MHz). It is infested with patents (over 250). EPC Global IS 1.0 was published in Sept 2005. It's a land of proprietary systems (IBM, Sun, SAP, offer full infrastructure). Several countries develop their own specifications: Japan UID, China RFID Group, etc. Software standards are to be developped.
Today, Wal-Mart/DoD top 100 suppliers are shipping boxes and pallets tagged with RFID. RFID is not integrated into production. Data is transported over emails. So far, we're only talking of limited pilot projects. Yet, VCs already invested $1 billion.
Several ObjectWeb projects are available that would make sense in a software infrastructure for RFID: SensorBean, OSCAR: Reader Interface; Sensor Bean: reader management; JORAM, ProActive, XQuare, OCtopus, C-JDBC: Data Filtering, transport; JOnAS, Celtix, Petals: Data Request/Legacy System Support; SNAP, ProActive: RFID Information Network.
Some countries (China to begin with) already work on a counter proposal to EPC Global's ONS. ONS is based on DNS. The problems are: it's centralized administration and control; it's controlled by the US alone; it relies on 40 year old architectures. This is an opportunity for research in Grid data and overlay network.
The goal of a RFID initiative at ObjectWeb would be to reuse existing components, attract new partners, foster OS/Standard organization cooperation and eventually start an ecosystem in this field. EPC ALE/IS is the only emerging standard in the RFID middleware area, with certification/test suite not available till Q206. The ObjectWeb board accepted the proposal for starting a new initiative in October 2005. Initial participants may include: FirstOpen, GMRC, Macnica, ScalAgent, SensorBea, Yangfan Soft. Tangfan is one of the 54 members in the Chinese RFID standard workgroup under Chinese Ministry of Information Industry.
RFID (Radio Frequency IDentification) are small tags able to transmit unique ID through radio. They are expected to replace Bar Codes in the 10 years to come. Mandates from Wal-Mart, DOD, Target, Metro is a driver for adoption of RFID and next wave of IT business. A RFID can be attached to virtually anything, including human beings. Nokia recently introduced the RFID scanner cell phone. An application may be buying goods in vending machines while having their price imputed directly on your phone bill.
A vision of the future of RFID is an "Internet of Things"; tags everywhere, readers everywhere. It would be necessary to leverage the Internet to carry data. Current RFID tags (namely the EPC class of tags) have limited capacity of 96/128 bits to store their (unique) id. Providing a network infrastructure falls into the scope of ObjectWeb.
Exemples of applications include tags on boxes, tags in passports, fight against counterfeit drugs, livestock, pets, kids... Sounds to me that nobody wonders whether this future would be a Brave New World or just... hell.
Gartner's Hype Curve analysis puts RFID Peak of Inflated Expectations in 2004, Trough of Desillusionment in 2006 and Plateau of Productivity in 2018. Market size in 2008 is expected to be around $6.5 billion. The big players are the usual suspects: IBM, Microsoft, Sun, Oracle, BEA and RadioActive & FirstOpen in open source. In the Java space, two JSRs have been proposed by Nokia about the reader interface.
In the architecture, data filter and transport is an important aspect, because unlike bar codes, RFIDs are constantly on and the readers keep reading information. This is somewhat similar to what happens with mobile (cell) phones. Management of RFID infrastructure extends beyond managing the reader. For instance, it would also include management of sensors to detect moting of packs/trucks transporting the goods. The information network is necessary to share information in and between organizations.
EPCGlobal is leading the current specifications. It's mainly a hardware specification, American centric (generation 1 uses 900MHz). It is infested with patents (over 250). EPC Global IS 1.0 was published in Sept 2005. It's a land of proprietary systems (IBM, Sun, SAP, offer full infrastructure). Several countries develop their own specifications: Japan UID, China RFID Group, etc. Software standards are to be developped.
Today, Wal-Mart/DoD top 100 suppliers are shipping boxes and pallets tagged with RFID. RFID is not integrated into production. Data is transported over emails. So far, we're only talking of limited pilot projects. Yet, VCs already invested $1 billion.
Several ObjectWeb projects are available that would make sense in a software infrastructure for RFID: SensorBean, OSCAR: Reader Interface; Sensor Bean: reader management; JORAM, ProActive, XQuare, OCtopus, C-JDBC: Data Filtering, transport; JOnAS, Celtix, Petals: Data Request/Legacy System Support; SNAP, ProActive: RFID Information Network.
Some countries (China to begin with) already work on a counter proposal to EPC Global's ONS. ONS is based on DNS. The problems are: it's centralized administration and control; it's controlled by the US alone; it relies on 40 year old architectures. This is an opportunity for research in Grid data and overlay network.
The goal of a RFID initiative at ObjectWeb would be to reuse existing components, attract new partners, foster OS/Standard organization cooperation and eventually start an ecosystem in this field. EPC ALE/IS is the only emerging standard in the RFID middleware area, with certification/test suite not available till Q206. The ObjectWeb board accepted the proposal for starting a new initiative in October 2005. Initial participants may include: FirstOpen, GMRC, Macnica, ScalAgent, SensorBea, Yangfan Soft. Tangfan is one of the 54 members in the Chinese RFID standard workgroup under Chinese Ministry of Information Industry.
posted by Francois Letellier @ 2:57 PM
0 Comments:
Post a Comment
<< Home