Observations from the IEEE 802 Plenary Session

March 18th was the student open day of the IEEE 802 plenary session held in Beijing. I was invited by MSRA to attend. The participants in the standard-setting process are all professionals, and I was basically like Granny Liu visiting the Grand View Garden, just there for the amusement. Since photography and recording were prohibited at the venue, and the technical documents discussed at the meeting were not public, there are no pictures or solid evidence to share.

First, let me explain what IEEE 802 is. IEEE 802 is a committee under IEEE (Institute of Electrical and Electronics Engineers), responsible for the establishment of local area network and metropolitan area network standards. The physical layer and link layer protocols of computer networks are basically established by this organization. IEEE 802 holds three plenary sessions each year, most of which are held in North America. Voting rights are granted from the third participation in the plenary session.

IEEE 802 has several working groups, for example, 802.3 is responsible for Ethernet, which is the wired network we use; 802.11 is responsible for Wireless Local Area Network (WLAN), commonly known as wifi. Each working group still has a lot to do. For example, Ethernet has 100M, 1G, 10G, 40G, 100G, and the 400G under research. Not only are the speeds different, but the transmission media used are also different; WLAN has 802.11a/b/g/n/ac/ad standards, not only are the speeds different, but the frequency bands used are also different. Therefore, each working group has Task Forces and Study Groups.

How Standards Are Born

The birth of a new network physical layer and link layer standard takes several years:

1. Someone proposes an interest (Call for Interest), and a Study Group is established in the Working Group. For example, the 400G Ethernet is currently in the “Study Group” stage. The meetings of the Study Group are quite similar to academic conferences. Everyone needs to prepare slides, speak for several tens of minutes, and interact with the attendees. These slides can be irrelevant or even conflicting, do not necessarily propose a complete solution, and do not need to reach a consensus, so this is a brainstorming stage.
2. After the Study Group comes up with a plan that everyone basically agrees with through lengthy negotiations, a Project Authorization Request (PAR) is formed and submitted to the Working Group for review.
3. The Working Group establishes a Task Force to standardize this plan. This is basically a process of everyone arguing and nitpicking over words. After revision after revision, a draft standard is produced.
4. The draft standard is submitted to the Working Group for voting. If it receives more than three-quarters of the votes, it enters the review stage; otherwise, it is sent back for rewriting.
5. The review is basically an administrative process. For example, the frequency bands and power used by wireless protocols need to be negotiated by the Technical Advisory Group within the Working Group and government departments such as the US FCC and the Chinese Radio Administration.
6. Finally, it is voted on at the IEEE 802 plenary session and becomes an official standard. Since each standard has its own professional field, members of other working groups may not understand its technical details. Plus, these standards have been carefully worded, so the pass rate of the plenary session is predictable.
7. After the official standard is released, the Task Force does not disband, but proposes amendments to the standard over time to solve problems encountered in the application of the standard.
   If you wait until the standard is released to start developing products, you may already be several years behind others. Therefore, large companies are eager to send their employees into the various working groups of IEEE 802 to participate in standard setting. The game between all parties is mainly reflected in the process of the Task Force arguing and nitpicking over words. On the surface, the meeting seems long and boring, and I, a non-professional, can’t understand it at all. In fact, the wording of the standard is very particular.

For example, there is a world of difference between may and shall: if may is used, it means that other companies can adopt different implementations, which can lead to incompatibility between different products (is this rare in the browser field?); if shall is used, it means that there will be no incompatibility between products in this regard, but the company whose proposal is adopted can seize the market opportunity and may even profit from related patents.

Open Standards and Innovation

The first thing in the morning was the chairman of the 802.3 working group and the vice chairman of the 802.11 working group (I hope I remember correctly), the two most authoritative working groups of IEEE 802, introducing the standard-setting process of IEEE 802 to us students. The first thing they talked about was the significance of Open Standard. When it comes to standards, the first thing people might think of is rules and regulations. No one wants to be constrained by these rules and regulations, so what is the significance of standards?

“Open Standard” is composed of two words, “open” and “standard”. The meaning of “standard” lies in interoperability, and “open” prevents users from being locked into a particular vendor’s product (no lock-in). Open standards build the infrastructure for innovation, allowing various innovations to coexist and benefit each other in a compatible ecological environment. The telephone network lacks openness, so it does not have the vitality of the Internet; various BSD derivatives lack standards, so they are not as popular as Linux.

In the process of “nitpicking” in standard setting, may and shall reflect the contradiction between the two words “open standard”: may means open, shall means standard. All companies hope to use “standard” to strengthen their established market position and use “open” to win a place for newcomers. A lot of compromises need to be made for an open standard to ensure that things made according to the standard can achieve certain functions, and leave enough space for innovation in usage, performance, and power consumption improvements.

Venue

The conference was sponsored by Huawei and held in a five-star hotel. Each signboard had the IEEE 802 and Huawei logos, as well as a QR code for viewing the agenda online. The venue can be described as luxurious, spacious and comfortable. I didn’t attend the opening and closing ceremonies, but rather the meetings of various working groups. Dozens of research groups or task groups each had a meeting room, and the venue occupied two floors in two buildings, which was quite extravagant.

Because we didn’t look at the map and schedule, we students wandered around like visiting a zoo, entering a meeting room, listening for two minutes and finding it incomprehensible, then leaving in disappointment.

“Chanting Group”

Most of the meeting rooms were engaged in “word-picking” work. A chair and two vice-chairs sat on the podium, reading the standard draft or amendment text. The attendees could also see these texts on their laptops, but of course, we students didn’t have this privilege. When someone had an objection, they had to stand in front of the podium, in the middle of the venue, and express their opinion at the microphone. Only the people on the podium could discuss with the person at the microphone. If someone had a new opinion while someone else was speaking, they had to line up behind that person and speak in turn. That is to say, at any time, only the three people on the podium and the person at the microphone could discuss, and others could not speak. Sometimes there were three or four people lined up behind the microphone, and sometimes no one had any opinions, and the podium seemed like a chanting session.

Regardless of whether the discussion was active or not, since I didn’t understand the background of the standards and couldn’t see the standard text, their discussions seemed like chanting to me, so I called these task groups that agreed on protocol details the “Chanting Group”.

One “Chanting Group” worth mentioning is 802.22 WRAN (Wireless Regional Area Networks). The chair of that group was Indian, but he spoke very standard English. Since we arrived at that group during the tea break, the chair kindly let us introduce ourselves and pulled out a popular science slide to explain to us. 802.22 WRAN uses the blank area of the television spectrum for communication. Due to the long-distance transmission characteristics of this frequency band, it can achieve long-distance network communication in the suburbs at a relatively low cost. This technology can be used in smart grids, IP TV (on-demand television), suburban surveillance, and other fields. I also asked a question, what is the difference between 802.22 and Google Loon, he answered that, first, Google Loon also needs a certain frequency band to receive signals and relay between balloons, and second, Google Loon requires a certain cost and cannot cover the entire surface of the earth, so 802.22 and Google Loon will complement each other.

“Report Group”

The other part of the meeting rooms were giving presentations similar to academic conferences, of which I could understand some. These should be research groups (Study Group), where each party proposed their own solutions and experimental results.

Among the several presentations I attended, I was particularly interested in the afternoon 802.1 TSN (Time Sensitive Network) meeting. When we entered, a person from Broadcom was talking about their solution. Seeing the familiar slide template, I knew this was something about data centers that I could understand. I listened to the entire Cisco presentation. From the beginning of the presentation, someone from an unknown company kept “finding faults”, and the presenter kept saying “let’s continue”. I originally thought he was at a loss for words, but all the questions raised were addressed and resolved in the later slides. After the presentation, no one questioned it anymore, it seems that Cisco is still the leader in the field of network switching equipment.

The presentations I didn’t understand were basically made by Chinese people, and their English was not as easy to understand as foreigners’. One person didn’t make slides, but used an Excel table to present in front of hundreds of people, and I couldn’t see clearly even sitting in the third row. I guess he had prepared his speech somewhere and read it word by word like reading a standard text. I, who is very poor in English, couldn’t understand at all. The official attendees had the electronic version of this Excel table, and they were all staring at their laptops; we students, after listening to the incomprehensible speech for more than ten minutes, collectively fled the venue. It is said that there were a lot of valuable information in the presentations afterwards, but unfortunately, I didn’t hear them.

After a day, the biggest gain was to realize the importance of listening and speaking. Reading documents in black and white and communicating with people are completely different things!