Company Offers Customizable Web Spidering 46
TechReviewAl writes "A company called 80legs has come up with an interesting new web business model: customized, on-demand web spidering. The company sells access to its spidering system, charging $2 for every million pages crawled, plus a fee of three cents per hour of processing used. The idea is to offer Web startups a way to build their own web indexes without requiring huge server farms. 'Many startups struggle to find the funding needed to build large data centers, but that's not the approach 80legs took to construct its Web crawling infrastructure. The company instead runs its software on a distributed network of personal computers, much like the ones used for projects such as SETI@home. The distributed computing network is put together by Plura Processing, which rents it to 80legs. Plura gets computer users to supply unused processing power in exchange for access to games, donations to charities, and other rewards.'"
Nifty... (Score:3, Interesting)
Who are the customers? (Score:4, Interesting)
I can see how they might get a fair number of people to donate their spare cycles for this, if the rewards are seen as sufficiently interesting. But are there really a whole bunch of startups (or other companies) that are really champing at the bit to create a new search engine? Other than marketers or malware purveyers, I mean. And do these searches honor robots.txt exclusions?
BTW I took a quick look at 80legs' website in an attempt to get these answers. I came up empty in that regard - so I will comment on how the CEO's hair makes him look like an in-disguise member of the Conehead family. Seriously, what's with the hair?
I'm a little confused (Score:3, Interesting)
Is there really a big demand out there for outsourced spidering? I had not heard of this market. They seem to be implying that there are all these start-up outfits out there who have invented really amazing, unique UIs that allow people to find exactly what they need on the Web, and all they need to be successful is access to a searchable index. Huh??
I mean, if you're going to be some kind of start-up search engine or "semantic company" (whatever that means), shouldn't Web spidering be your core competency? If you're going to differentiate yourself in the market, how can you buy spidering as a commodity? How to you expect to attract any investment if you're telling potential investors that you rent your spidering capability from another start-up -- let alone one that uses some kind of half-baked P2P technology to do the work?
Seriously, in a world where Google seems willing to partner with just about anybody who needs any kind of searching for reasonable rates, what is this company's proposed customer base? (And no, the Technology Review article includes no quotes from customers at all.)
Rent our botnet! (Score:3, Interesting)
This looks like an attempt to monetize a botnet. What, exactly, do the people running their "client" get out of this? Do they know they're sucking bandwidth, and possibly being billed for it, on behalf of someone else?
I run a web spider [sitetruth.com] of sorts. And I know the people who run a big search engine. Reading the web sites isn't the bottleneck. Analyzing the results and building the database is. Outsourcing the reading part doesn't buy you much. If this just did a crawl, it would be of very limited value. That's not what it does.
What they're really doing [pbworks.com] is offering a service that lets their customers run the customer's Java code on other people's machines in the botnet. That's worrisome. There are some security limits, which might even work. Supposedly, all the Java apps can do is look at crawled pages and phone results home. Right.
This thing uses the Plura botnet. [pluraprocessing.com] "Plura® is a grid computing system. We contract with affiliates, who are owners of web pages, software, and other services, to distribute our grid computing code. We utilize the excess resources of peripheral computers that are browsing the internet when such browsing leads to a web page of one of our affiliates. That web page has imbedded code that allows the visitor to participate in the grid computing process. We also utilize embedded code in software and other services to allow such participation." Not good.
The main infection vector is apparently the Digsby chat client [lifehacker.com], which comes bundled with various crapware. The Digsby feature list [digsby.com] does not mention that Plura is in their package.
This thing needs to be treated as hostile code by firewalls and virus scanners.
Re:Rent our botnet! (Score:2, Interesting)
Outsourcing the reading part doesn't buy you much. If this just did a crawl, it would be of very limited value. That's not what it does.
Wrong. If I want to spider a single web site, many sites have rate-limiters that kick in and will block me after a while. This would allow me to hit it from multiple machines.
There are some security limits, which might even work. Supposedly, all the Java apps can do is look at crawled pages and phone results home. Right.
Why the sarcasm? This seems like a perfect use case for the JVM's security mechanism.