Last week MSR Cambridge held a press day, in which we talked about a number of our research projects, ranging from proofs of the 4-color theorem to new kinds of user interfaces. One of the projects we discussed was Avalanche, an early research effort looking at peer-to-peer (aka P2P) distribution networks.
MSR has done research on peer-to-peer networks for years, and published lots of papers on various aspects of the topic. From a technical point of view, they can have lots of advantages over other architectures, including load-balancing, removing single points of failure, reducing latency, and improving availability -- if architected correctly, of course. P2P architectures can potentially be applied to a number of different application areas, including software and media distribution.
The funny thing about putting on an open house for the press, is that you can't always predict what the one thing is that everyone will want to write about. This time, it turns out that it was Avalanche. Why? One word: BitTorrent. (I guess that's one word)
Here's a little bit more background on Avalanche. One of the potential problems with P2P distribution systems is the "last byte" problem. If your scheme is to break up the original file into smaller pieces and distribute them around, that means that recipients need to do the reverse: identify at least one copy of each of the N pieces, potentially spread across N machines, download them all, and stictch them back together. If N-1 of those machines are fast and responsive, but one is under heavy load or has a bad network connection, then you're not done until that one slow machine finishes sending its piece of the file. Further, if one of the pieces suffers from data corruption, then the end result will be corrupted. The researchers on the Avalanche project are working on ways to avoid the "last byte" problem, and to provide some error-checking ability. Somewhat independently, but equally important, they are looking at how to support the notion of "authorized" content, which could be an official, verifiable patch to a piece of software, or a way to protect the intellectual property rights of content creators.
Last week the researchers described these broad parameters of their research, and drilled down more on the network protocol work. They also released a research paper that went into more detail on their algorithm and in order to evaluate the efficiency of their algorithm, provided benchmark comparisons they had done to the most established P2P distribution system out there today: BitTorrent.
Now, keep in mind: this is Microsoft Research doing this work. We're a research lab, not a product group. 700 people worldwide, doing research in about 55 different areas of computer science. Publishing papers, and yes, transferring technologies to Microsoft product groups (as well as licensing externally). But MSR doesn't build and release products. Even if we did, the Avalanche research project isn't a product. It's a bunch of core technologies, and we've published a paper so that the larger research community (and the field as a whole) can review it, critique it, and potentially choose to adopt or improve it further. But MSR doesn't ship products.
Of the reporters who were there, nearly all of them got the story right -- this is research, it's interesting, and the comparisons with BitTorrent performance are apt. Also, there's no commitment or timetable for MS to release Avalanche in any form. Here's one good article, other than the part about "beta testers" which isn't true -- we have no beta version (it isn't a product), and there certainly aren't beta testers. Here's another, from none other than the Register. But one article overstated parts of the story, for a nice dramatic effect. Personally, I would love it if MSR research projects actually had the ability to directly influence the stock price, but alas, I must leave that to reporters' imaginations.
Late last week, we started seeing the second-order stories; starting, of course, with Slashdot, which despite the inflammatory headline overall was pretty rational and balanced in its discussion, with several people pointing out that this really is just a research project focused on some specific technologies. But other reporters picked up and re-reported based upon the first round of coverage (or simply reposted), and the fun began in earnest (I'm being sarcastic here). Of course, it spread like wildfire through the blogosphere, with the usual decreasing accuracy and increasingly angry headlines as this morphed into an attempt by Microsoft to crush BitTorrent.
By early this week we were on the third-round over coverage, and the hype had really taken off.
Today was the kicker: Bram Cohen denouncing Avalanche as "vaporware" (as part of a broad tirade) and John Dvorak declaring that Avalanche is part of a Microsoft consipracy to discredit BitTorrent. I like Slashdot's take on Dvorak's little rant, and just for completeness' sake, here's their take on Cohen's commentary.
Um, let me get this straight. In six days, a research project went from some algorithms in a paper to Microsoft's competitive answer to BitTorrent, to "vaporware" to an evil conspiracy.
Oh please. We've never claimed that this was anything more than a research project. We released a paper so that everyone -- including Microsoft's fiercest competitors -- knew exactly what we were doing, and could comment. On that point, Bram Cohen's technical comments about Avalanche are useful and appreciated, especially in the comparison to BitTorrent's algorithms which he obviously is the world expert on. The research community is built upon healthy skepticism; work should be scrutinized, replicated, and subject to discussion. We're glad for that.
MSR, and Microsoft as a whole, believe that research should be done on existing technologies (including BitTorrent) to find ways to improve them. Everyone benefits from that. We also believe -- though we recognize that others don't -- that content owners should have the ability to protect their content in appropriate ways (they may choose not to use that ability), and if we can create technologies that combine the best network efficiencies of P2P networks with ways to appropriately protect content, great things could happen. I'm not going to go into the whole "fair use" issue, because it's been debated widely in endless fora, and because P2P technologies and IP protection are not just about music and movies -- there are other uses for a system like this, including making sure that software patches are officially authorized and don't contain malware. I will say, though, that Microsoft and MSR are very supportive of open debate and discussion on content protection and fair use. We've even hosted Cory Doctorow here at MSR to give a (now famous) talk and share his views on the topic (which we don't necessarily agree with).
Hopefully this sets the record straight. Now can we please stop the press/blog feeding frenzy and get out of the reality distortion field on Avalanche?