Features

Mesh creator to be overturned by the big guys?

There is one Mesh technology which has been creating (commercial) neighbourhood networks for 18 months now; a new Mesh is set up somewhere in the world every day using it, and it is fully WiFi compatible. Intel and Cisco have dismissed it as being "non-standard."

Back in March, Intel announced its own Mesh technology. At the time, this reporter said: "It looks as if Intel has decided to ignore everything that has been done in linking wireless network nodes into a mesh - and has, nonetheless, impressed observers with its originality of research!"

It seems that in six months, Intel's research, like Cisco's, has advanced not at all.

The two wireless giants are planning to rescue the world from the chaos, which they appear to believe, lurks around the corner. But magically, neither Intel nor Cisco has apparently heard of LocustWorld. At least, they don't admit to in in public.

Since the announcement (at last week's WiFi Planet show in San Jose) that "The two Silicon Valley-based companies said that they will introduce an industry standard for wireless mesh networking during an inaugural study group at the Institute of Electrical and Electronics Engineers (IEEE) meeting in Vancouver, Canada next month," reported in Wi-Fi Planet itself there have been - justified - cries of outrage.

Strangely, however, the cries of outrage have centred on the lack of recognition for some very obscure rivals. They have mentioned FireTide - which has yet to ship a single mesh, and PacketHop which is building a network already! - in preparation for a trial next year, and BelAir and Ember.

BelAir "introduced" its first product in October this year (if it has shipped any, it's being very modest in not admitting this publicly) and Ember is concerned with the still very futuristic technology called Ultra Wide Band. Nobody is planning to roll out LANs using UWB.

So, with the LocustWorld Mesh well over a year old, tried and tested and in full time commercial use all over the world, two questions arise. First, why do Cisco and Intel need to come up with a new standard? and second, why are they steadfastly ignoring the one de facto standard with a track record?

A cynic would look at the track records of Intel and Cisco. They are all for de facto standards - when they invent them, that is. Intel is very proud of the de facto x86 processor architecture, for which it never sought IEEE approval, and Cisco will tell you till you're short of breath, how wonderful its LEAP security extensions to WiFi are. Cisco is so proud of LEAP that it has embedded the technology in the CCX extensions which it is licensing to any maker of WiFi clients, with the object of making sure that it's a de facto standard which nobody else can offer - because it isn't offering it to other makers of access points. And yet the IEEE has introduced the 802.1x security standard, which isn't part of CCX ... what happened to Cisco's enthusiasm for IEEE standards, there?

Naturally, when the folks at LocustWorld heard about the proposed new standard, they felt the rumble of a juggernaut. So they contacted Intel, and asked whether they were deliberately ignoring them, or just had, somehow, managed to avoid hearing about them.

The answer, apparently, was along the lines of "Yes, we have heard of you. But you aren't a serious commercial offering, are you?"

One would imagine that people doing seminal research in a field would have the keenest eye for what their rivals are up to. Indeed, normally, this is the case: whenever I interview someone about a "first in the field" technology, they have always heard of all their potential rivals, and are an excellent source of information about "what can go wrong with the other folk."

Frankly, it beggars belief that Intel's researchers can have heard about the MeshBox without discovering that it is in fulltime, commercial production. The LocustWorld Mesh design has recently been rolled out in Gun Barrel City in Texas; in Louisiana as well as in Malaysia, the West Indies, and around Europe.

We recently reported on how one of Europe's largest betting operations, Tattersalls had become a LocustWorld customer in Newmarket, near Cambridge.

How is it that despite the fact that Google has 5,000 odd hits for LocustWorld, that the technology has been mentioned not just on specialist little sites like this one but also on the BBC site and dozens of Linux specialist publications, that Cisco and Intel can complain that there is "no standard"?

You get a clue, if you read the original report from WiFi Planet. The argument is not one which is attributed directly to any named executive or engineer in either Cisco or Intel. Instead, the site quotes one of its own staff: "If you put all the mesh network boxes in a room and turn them on, they do not talk to each other."

And, he goes on, "Some use cellular, some use 802.11a. Cellular is good for mounting on telephone poles for longer reach. 802.11a is good for shorter distances and is great because it won't interfere with 802.11b or 802.11g networks."

This is the reddest red herring.

There is nothing at all anybody can do, anywhere, to make cellular radios receive WiFi transmissions, or vice versa. No standard can reconcile 1900 MHz with 2.4 GHz and nobody is ever going to produce an 802.11a system which broadcasts or receives on 900 MHz. A mesh of wireless devices has to be a mesh which uses the same frequency spectrum, or else the problems of configuring a link from A to B to C will be impossible if C dies and D comes online, on another frequency. The radio layer has nothing at all to do with the way a mesh configures itself, or how it routes messages, or how it manages security.

Since neither kindly giant has put its name to a proposal, it's hard to know what they are thinking. But if the problem is the one described in the report, then either they are not thinking, or they are thinking of doing something rather different from offering a helping hand to a confused world.

What they are probably thinking about, is WiMAX. At the moment, WiMAX is 802.16, and about to emerge in its latest variant, 802.16e - the magic "e" being the one that allows WiMAX to cope with roaming.

Originally, WiMAX was a metro area networking technology, using fixed radios. It has an awful lot of advantages over WiFi, not least in its range - it's legal (in America, at least) to send the signal as much as 50 km where WiFi (in Europe, certainly) is restricted to low-power transmissions that normally cover only a few hundred yards. And it is also faster; you can provide broadband to more people at a higher rate.

Intel has revealed that it intends to build this new, mobile version of WiMAX into all its silicon. Every chip that Intel produces, will have a little 802.16e wireless built into a corner. It would make all the sense in the world for this to be turned into the cornerstone of a mesh technology so that any individual transmitter could reach others, which would forward messages to the next one, without having to dive into a fibre Internet link first.

The problem, of course, is that if there is already a de facto, working standard for this, then Intel wouldn't be able to charge royalties on that standard. Indeed, it might actually find itself in the position of having to pay royalties to someone else.

And if the de facto standard is based on LocustWorld, then Cisco's CCX standard becomes irrelevant.

And, of course, if all the wireless sites are primed to mention BelAir and FireTide and PacketHop and Ember, none of which are even shipping, never mind commercially viable, then it can hardly be a surprise if, in January, the IEEE (which has, of course, many engineer members who work for Cisco and for Intel) decides to ignore such work-in-progress efforts, however well intended. And, when it comes down to the wire, of course, a standard that was developed in the US by US owned companies, can only be far more suitable, in every way, than something concocted in some offshore area by someone who might not even be an IEEE member ...

Locustworld's technical mastermind, Jon Anderson, is planning to attend this inaugural IEEE meeting next month. It will be interesting to see whether he gets a fair hearing.