Slashdot Mirror
Akamai Wins Lawsuit to Protect Obvious Patent
brandaman writes "Akamai, the largest content delivery network (CDN) with about 70% market share, recently won its lawsuit against the against second largest CDN - Limelight Networks. The suit asserted that Limelight was infringing on Akamai's patent which, upon examination, seems to be somewhat on the obvious side. 'In accordance with the invention, however, a base HTML document portion of a Web page is served from the Content Provider's site while one or more embedded objects for the page are served from the hosting servers, preferably, those hosting servers near the client machine. By serving the base HTML document from the Content Provider's site, the Content Provider maintains control over the content.' Limelight is obviously not pleased, and this is not the first lawsuit Akamai has won regarding its patents."
You're saying that rewriting urls in a web page to fetch objects from geographically different servers was obvious in late 1998?
Yes its obvious. I remember discussing this in an ISP operators mailing list over 10 years ago.
Well, since HTML, circa 1993, was designed to allow for referencing external components. Rewriting URLs was a fundamental principal for serving pages and applications, which existed since at least 1996 that I'm aware of, as I did it then. Add in to that that commercial IP blocks are owned by companies with definite locality, and I'm not sure what part of Akamai's patent isn't stating things that were already in existence.
It was already stated that algorithms cannot be patented. And that's all that Akamai seems to have. Not withstanding that there's several thousand people skilled in the arts that would have come up with the same or similar solutions to the problem set.
The cesspool just got a check and balance.
It's a combination of modified URL's in the static HDML, DNS trickery that causes those URL's to be downloaded from servers physically close to you, and smart caching of that content. It basically provides a way of ensuring that static content like images, which take up a lot of bandwidth compared to HTML documents, is downloaded from servers physically near you and not from the companies primary server. It dramatically speeds up the loading of web pages no matter where the requests come from, and offloads a lot of processing & network utilization from the site serving up the HTML.
I posted more details in a reply here if you want a little more insight into how they do it.
There are two parts to a non-obvious invention. Either it has to be a solution to a problem that is not obviously a problem, or it has to be a solution to a (potentially obvious) problem that isn't something anyone would think of. The problem that HTML pages change a lot and big binaries don't but take a lot of bandwidth is an obvious problem. The solution of caching the binary bits near the client is an obvious solution.
I am TheRaven on Soylent News
I would assume they are talking about Edge Side Includes and not simply about the serving of images.
ESI is like Server Side Includes, except that the included part resides on the Edge servers. So your server would serve a page with only the content personalized to you specifically (like the fact that you are logged in) but a box full of news headlines that everyone sees would be included by the edge server.
Not entirely obvious, but I am not so sure it warrants patent protection in any case.
Cutting boxes to minimize waste and facilitate processing can be solutions to VERY nonobvious problems easily desrving patents.
Let's start first with the definition of "obviousness." Patent law doesn't go by the common English or Webster's definitions of the term - it has a very technical meaning, refined by probably a thousand patent law cases. Many nuances. And unhelpfully, the definitive section on the topic is circularly defined.
At least two reasons for the technicalities. First, virtually anything is "obvious" in hindsight. Second, what is "obvious" to one (unbiased) person may be completely non-obvious to another (unbiased) person, and the patent office would produce radically inconsistent results if examiners were permitted such subjectivity.
- David Stein
Computer over. Virus = very yes.
While I haven't been following the litigation very closely, Limelight's position on it seems interesting. They seem to be denying that they infringed upon the patent, rather than taking the approach that the patent is obvious or that there exists prior art. I'm sure there is a sound legal strategy here, I'm not a lawyer - but my instinct would be to go the other way (maybe that's why I'm not a lawyer, ha!)
It'll also be very interesting to see how the other well-funded CDN players react to this - Level 3 (which is also suing Limelight over patents), EdgeCast (Disney-funded) and Internap (with about the same market cap as Limelight.)
The way I read the patent, they must all be infringing in some way, no?
... automatic redirection to the "nearest" mirror?
Brilliant!
What a novel use of technology.
Surely this is just amazing. Who ever woulda thunk that computers could do things for us automatically?
If you disagree with me on social issues, then it's pretty clear that you are a narrow-minded bigot.
I was doing web caching at the time (I had my hands very early on on the original hotspots paper by Akamai's founders). When I learned of the embedded elements redirection I found the Akamai idea totally non-obvious and far more reaching in terms of web caching than their hotspots contribution. Of course, once I saw it, all I could say was "what didn't I think of that, its so obvious!"
Well, it was pretty obvious then, at least to people in the business, especially considering that at least one earlier CDN patent (e.g., US Pat. 5,991,809, originally filed as provisional pat. 60/022,598, filed on July 25, 1996 , by me) had already been granted and therefore made completely public in 1997. Clearway Technologies (my company) was already selling a commercial off-the-shelf CDN implementation system starting in September of 1996. Akamai's success has been substantial, and I feel it is truly well-deserved, but they were not the first to invent a CDN, nor the first to patent it, nor the first to bring it to commercialize it.
-Mark Kriegsman
Founder, Clearway Technologies
There may be all sorts of trickery involved in what makes Akamai work, but the patent covers more than that trickery. It covers any system where a webserver modifies a URL to include a hostname whose DNS entry is served up by two DNS servers in the system, and whose content is served up by a host other than the webserver.
For examle, if I have a webserver at example.com, and it modifies image URLs within it to point to foo.bar.example.com, and there's an 'example.com' DNS server which contains the NS record for 'bar.example.com', and a 'bar.example.com' which has an A record pointing 'foo.bar.example.com' to some machine other than example.com, I've infringed claim 1 of the patent. If there are two nameservers for example.com, I've also infringed claim 2. If there are two nameservers for bar.example.com, I've infringed claim 3.
The courts are accepting these so far.
Yeah, lower US courts decided to start allowing software patents.
However in the current Microsoft-AT&T case before the US Supreme Court the multiple justices were clearly skeptical of that behavior. In particular:
JUSTICE BREYER: I take it that we are operating under the assumption that software is patentable? We have never held that in this Court, have we?
MR. JOSEFFER [DOJ Atty]: No, but as I was saying before -
JUSTICE BREYER: So what should we do here? Should, if we are writing this, since it's never been held that it's patentable in this Court
I have read the Supreme Court rulings relating to software patents. The rulings were back in the early 80's or so, back before the lower courts went off on their software patent kick. It seems clear to me that the lower courts have ignored or directly violated several points of Supreme Court law in those rulings. It appears that the Supreme Court is looking to directly rule on the subject and reign in the aberrant lower courts.
-
- - You can't take something off the Internet! That's like trying to take pee out of a swimming pool.
Edge caching in general is obvious. The implementation is not, and that's what this lawsuit is about.
Limelight copied akamai's patented edge cache implementation, and violated enough of the patent to warrant this decision.
I can see how a bunch of people jump on the "obvious" bandwagon, developing an edge cache for a single enterprise would be relatively easy to do. Developing an edge cache infrastructure that will work across hundreds and thousands of different enterprises with different business and process infrastructures with varying and often conflicting traffic load patterns is an entirely different problem.
Let's look at the url rewriting aspect of it. The rewritten urls are specified to include popularity flag and use virtual hosts as a serial #. Something that would be obvious today, but in '98/'99 not so much.
Let's look at geographic dispersion. What is your obvious geographic dispersion methodology? A database like MaxMind has to offer? Those were few and far between in 1999. They use a network map. Further, they also consider the load of their servers when returning dns responses - so we're not just talking about getting the closest server to the edge, we're talking about getting the closest server with the lowest load at the edge. There's also the problem that the dns request is in most cases not coming from the client machine, but from a geographically disparate dns cache. The akamai system redirects users to a closer server if one exists and that situation is discovered by the web server.
There's also the situation that the network load of a particular server becomes too heavy while serving a particular request midstream (i.e. for audio and video). This patent covers switching the responsibility for handling a request to another server midstream.
This patent also includes anti-DOS/DDOS technology - a no brainer today, but in 1999? Not so much.
I wouldn't categorize Akamai as good or bad. I don't know much about the company itself. I can see the possibility of working around their patent to build a competitor - it's not an all-inclusive-only-we-can-do-CDN patent. But given the fact that limelight's foundation was built by a bunch of Akamai's ex-employees, it certainly isn't surprising that they chose the same path for resolving issues that Akamai did. And given Limelight's close ties with Microsoft, it's also not very surprising that they chose to emulate what they knew rather than innovating and improving upon the model.