SP100 is a new standard being developed for industrial wireless sensor networking. As part of the ISA, SP100 is one of many industrial standards emerging from the organization. Next week, there will be a meeting of SP100 members held in conjunction with ISA's annual conference, ISA Expo. In this article, I'll recap the current activities of SP100 prior to the meeting next week.

The first official SP100 meeting, a teleconference, was held on April 20, 2005. Two groups were formed--SP100.14 to standardize low power wireless sensor devices and SP100.11 to standardize higher bandwidth wireless infrastructure and backbones. The work accomplished by the group from April 2005 through July 2006 culminated in two major documents, the SP100.14 Request for Proposals and the SP100.11 Request for Proposals.

Notification of intent to submit proposals was due in late August, and proposers presented slides on the ideas behind their proposals at a meeting in Raleigh, NC from September 6-8. At this meeting, the following companies presented their ideas for SP100 (some only attacked a single layer of the OSI model, while others addressed all layers): Analog Devices, Adaptive Instruments, AIN, Apprion, Certicom, Crossbow, Dust Networks, Emerson, GE Global Research, Machine Talker, Nanotron, Newtrax, OMNEX, Oak Ridge National Laboratories, Sensicast, Siemens, ST, Texas Instruments, Honeywell, and Yokogawa.

Although such a large number of companies proposed methods for the SP100 specification, many of the ideas were extremely similar. Over the past month since this meeting, the proposers have aligned themselves into a few major groups. Below I'll detail each of these groups and the key points of their proposals. Keep in mind that these observations are my personal opinion--please contact me with any omissions or corrections.

WNSIA: Otherwise known as "Wireless Network for Secure Industrial Applications". The group is led by Honeywell with members including 3eTI, Adaptive Instruments, Crossbow, Endress+Hauser, Flowserve, Omnex Controls, and Yokogawa. WNSIA proposes a primary network of 802.11 based routers with wireless sensors communicating with routers in a star topology. Both the mesh 802.11 backbone and wireless sensors communicate using TDMA schedules. Of note is WNSIA's proposal to use proprietary radios for the wireless sensors; specifically narrow band 2.4GHz frequency hoppers. This PHY proposal has been called into question by other members of SP100 on the basis of interoperability and coexistence, two items that all IEEE radios must adhere to for standardization.

Emerson/Dust: Emerson and Dust are proposing a full mesh wireless sensor solution based on Dust's TSMP (time synchronized mesh protocol). The protocol operates above the IEEE 802.15.4 PHY at 2.4GHz and assigns all communication into TDMA time slots. The scheduling is centrally controlled, so any change in the environment or location of a sensor requires a new schedule to be computed and disseminated from server to nodes. The Emerson/Dust proposal is the only proposal that actually uses a full-mesh architecture (albeit centrally controlled) instead of a star, tree, or cluster-tree architecture. A whitepaper on TSMP can be downloaded from Dust.

Sensicast/STG/GE: Sensicast's network architecture is similiar to Dust's, with the exception that they offload some of the work to be computed distributedly by the nodes. The system is set up as a cluster-tree, with router nodes communicating in a tree back to gateways. End nodes communicate with routers in a star topology, with communication slots distributedly assigned by the routers. Of primary note, in my opinion, are the concepts that STG brings to the group. STG proposes "sensor network tunnels", a way for sensor data to traverse one side of a sensor network to the other by first going to a gateway, tunneling through the IP infrastructure, and then popping out the other side back into the sensor network--an idea that utilizes the backbone network (which has power and provisioning) much more effectively. GE lends their security model to the group, which is the onerous of the proposals due to the fact that it does not require SP100 members to license any technology to adhere to the proposed security standard. On the other hand, GE's security is not fully distributed and is not as comprehensive as the proposal presented by Certicom.

Siemens: Okay, so Siemens isn't a group but rather their own single entity. The Siemens proposal has the best of all worlds--sometimes we might need the performance of a star topology and sometimes we might need the versatility of a mesh topology. The result is a proposal with "Advanced Performance" and "Advanced Mesh" modes for network operation. What is undefined in the proposal is how I can deploy a network with both characteristics; commissioning and deploying a hybrid performance/mesh network seems phenomenally useful in the harsh industrial monitoring environments, but this notion has not yet been resolved in presentations or whitepapers.

Summary: Despite the variety of architectures proposed, there is a few common themes that almost every entity proposed. These include:

• IEEE 802.15.4 2.4GHz PHY and MAC because IEEE has already done the work of making sure it works well with others.
  
• Adaptive Frequency Hopping (or AFH for short). All proposers included frequency hopping as part of their proposals, showing how important this technique is for achieving high reliability when using lossy wireless links.
  
• Both TDMA and CSMA/CA networks were proposed, but the compromise appears to be an IEEE 802.15.4 MAC with modifications to the GTS (guaranteed time slots) mechanism to support AFH from the previous bullet. This would provide for a short contention period followed by pre-assigned slotted communication. Almost all proposals included a TDMA element.
  
• No dominant security and authentication architecture has emerged. Proposals with significant technical merit require IP licensing, whereas proposals that use unlicensed technology are not as comprehensive for all aspects of authentication.

Comments or questions should go to joe at moteiv.com. See you in Houston.

<< FIRE in the news | Main | News from ISA Expo in Houston >>