Custom Search

Thursday, April 9, 2009

How did a Radio Technology get Full-Mark (Barcelona 2)

(I hope this post could add your Easter fun. It is your "fault" to take it seriously. )

Continua summit on Friday Mar 27. I called it "amusing", but somebody may feel frustrating.

The technical working group started with the low power radio selection process report. First it was ANT taking back their yes vote to the low power PAN scoring rules, which was finalized the day before, stating they were "faced with an non PAN bias and refused a real voice of change". It was an uncomfortable moment. Somebody else murmured something, seemingly supported the statement. Some Zigbee company guys chuckled. The meeting chair controlled the process and did not allow this to disturb the atmosphere. The presentation moved ahead. Soon after, the TWG members enthusiastically voted yes to approve the scoring sheet of LP-PAN and LP-LAN.


I sat there thinking that the ANT complaints may be true, because Zigbee companies do control the low power radio process. Philips has staff sitting in the leadership roles of use case group, end-to-end architecture sub-team and LPR sub-team. Proposals from Nokia (BT LE), Polar (BT LE) and Dynastream (ANT) could hardly get through. No time to ponder. The important, long expected and with high-hope self-evaluation started, this time in a reversed alphabetical sequence.

Zigbee first. It was the former Zigbee vice chair Jon Adams of Freescale. Again he highlighted that over 30 Continua member companies are also Zigbee Alliance members. (I had thought he might want to hide this fact. But as he pointed this out, he should have further pointed out that more than 1/3 of the LPR sub-team participants are Zigbee members) After praising the very successful demonstrations on Wednesday, he started his pitch on the newly announced Zigbee health care profile and how Zigbee can benefit the Health industry. Jon had all kind of reasons to be very confident going into self-scoring in the last.

One of the important step in the LPR radio selection is technology scoring. The LPR team has worked for about a year coming out of two scoring sheets, one for LP-PAN and one for LP-LAN. The scoring sheet has 10 categories. Each category has 10 points. The ten categories are

  1. Transport Desired
  2. Transport Adoption
  3. Type of SDO
  4. Implemented Globally
  5. Costs
  6. Interoperability Support
  7. Continua Technical Alignment
  8. Security
  9. Power Consumption
  10. Networking Capabilities
The first category "transport desired" will be scored based on a Continua TWG "informal" survey result. And in the last minute, the cost category was pulled out because some legal concerns. So a full mark of the self-evaluation will be 80 points.

Zigbee first scored on LP-LAN. They gave themselves 78 out of 80. I have several technical doubts about their scoring on power consumption and localization. But as the score was this high, I could not help feeling Jon was mercy to reduce two points on transport adoption. Jon may feel giving Zigbee a full mark was not appropriate in this situation.

That specific criteria is about how many "radio IC" in history have been sold and a technology could get full mark if the volume is greater than 20 million. I thought Zigbee could give themselves full mark for this criteria though the volume of Zigbee is not that high, but the
802.15.4 radio IC has definitely reached that plateau.

The scoring on LP-PAN ran even faster, as there are many common criteria between LP-LAN and LP-PAN. Jon took an additional point off on power consumption, as he claimed they "have many products working on coin cells on 15.4" but he would not give himself that point. But on the presentation, it was written "can support coin cell on both ends." If this is true, let's expect to see Zigbee or 802.15.4 enabled sports watch soon in the market, as well as Zigbee or 802.15.4 enabled heart rate monitors and foot pods. Another big question of mine was the claimed 10 month battery life on the given usage scenario. So Zigbee rated itself 77 out of 80 on LP-PAN

The next were Sensium and BodyLAN. After seeing the high scores of Zigbee, I just could not wait to see the self score of Bluetooth Low Energy. And I know the presenter was Robin Heydon of CSR , co-chair of BT LE SIG. I had read the funny squabble betten Robin and Bob Heile , Chair of Zibee, before.

But you had to give Sensium and BodyLAN the chances. Sensium said they only do PAN and scored themselves 40.5. The battery life on the aforementationed use case is only about 4 month as they presented. BodyLAN gave themselves 59.5 for PAN and 58.5 for LAN. For BodyLAN, the key message was the suggestion that Continua should open its ecosystem to accept all PAN and ULP PAN technologies at the high application layer.

Finally, Robin was on the stage. But we had to wait for a long time. His slides were not provided at the start of the meeting. At the moment, Robin came with his Mac and insisted on plugging
in directly. Being refused, he eventually deposited the deck on a USB drive. I have to admited, Robin is one the best technical presenters I have ever met. On some of his slides are just one large number, such as the current shipping volume of BT chips, the predicted price of BT LE chip in 2009, 2010, 2011. He disclosed some detailed how-to of BT LE that are not public available.

After his blurb, Robin started score-scoring. The first criteria was the sales volume in history, he claimed if Zigbee could give themselves a 5 points for Radio IC history volume, they could claim
10 based upon reaching an assumption that silicon was released on spec version 0.9 and they had backward compatibility. The audiences laughed. For those who do not know why people laughed, different versions of Zigbee specifications (-2004, -2006, -2007, and ZigbeePRO) may not be compatible. That is the unspoken blemish of Zigbee and Robin dared to makes joke of it publicly. As he gave BT LE 10 point on this one, he had no hesitation to continue score 10, 10 and 10, even though the notes on the presentation seemed not suggesting a 10 mark for all. BT LE got 80 out 80 on LP-PAN first, then without suspense, 80 out 80 on LP-LAN.

By then the meeting room were enlightened and I knew what he did before his turn. Robin was high, but unfortunately he was cut off at the end for going too long and too much time was wasted to set up his presentation.

Robin's presentation gave the rep from ANT a hard time. But in general, he managed to make the whole atmosphere back to the ground. He urged people to look at the actual use cases not
the interpretations of IC history, battery life and other fictitious criteria. I was quite surprised to see the battery life for the use case was only 7 months or so, even shorter than Zigbee. (Should I blieve the number given by Zigbee without proof?) On his slides of the self score were many icons of frustrated face. ANT gave themselves 68.5 on LP-PAN and 69 on LP-LAN.

I have no idea how a serious result could come out. Maybe as I said in my post on the day of this Continua event , a delay of the low power radio selection is the best choice. Or as the BodyLAN guy suggested, Continua should open the low power radio part of the ecosystem to all.

No comments:

Post a Comment