Hacker News Reader

Hacker News

Comment by sp332 | Hacker News Reader

Parent Full thread

sp332

How did Aaron get access to the for-pay articles (page 9)?

Also: nice going, Aaron! Drag research access into the 21st century, kicking and screaming!

Does anyone think it's odd that an Acer laptop could write these files to disk faster than JSTOR could serve them?

View on HN

"Does anyone think it's odd that an Acer laptop could write these files to disk faster than JSTOR could serve them?"

Nope. I bet the JSTOR servers are serving many concurrent requests. If he had the servers to himself then yes, that would be surprising.

FTA: This is more than one hundred times the number of downloads during the same period by all the legitimate MIT JSTOR users combined

It sounds like he was using the majority of their resources. But maybe those servers serve lots of places, and they only mentioned MIT as an example.

They do. Universities across Canada also subscribe to this.

Yeah, my college also had it. But the article made it sound like only MIT was using these servers.

Anyone on MITNet has access to JSTOR articles free of charge to the user. Similarly the ACM, IEEE, etc. all have agreements like this with major universities.

Oh, so some articles are available to buy, and the subscription includes those and more? I thought the subscription was the subset.

JSTOR offers packages and any articles not included in those packages can be purchased by individuals. Most choose instead to request it via interlibary loan and you'd have it within 24 hours usually.

Yeah, I was laughing about how an Acer laptop took down their service and did damage to their network. If I was JSTOR, I wouldn't prosecute just because it makes our company look ridiculous.

Because, of course, JSTOR plans for 100x usage spikes and performs neither logging nor accounting, so serving a file is just as simple for them as downloading it is for the client.

> Does anyone think it's odd that an Acer laptop could write these files to disk faster than JSTOR could serve them?

Why would that be odd? SATA = 3 Gbps throughput with minimal overhead, Ethernet = 1 Gbps with lots of overhead (IP headers, Ethernet headers, HTTP headers)

SATA = 3 Gbps throughput with minimal overhead

USB HD's top out about 200Mbit/s. JSTOR's RAID arrays should be able to drown his laptop without breaking a sweat.

It sounds like JSTOR's servers aren't really optimized for high article download rates, to the point that his one laptop accounted for a significant part of normal continental US load. They probably had some bottleneck in the system that they never noticed before — maybe their logging infrastructure was absurdly slow or something.

JSTOR isn't at MIT; it's an online service. He just needed an MIT IP address to be able to access it without paying.

Did it specify that he was writing to local disk? If I did a stunt like this, I'd probably try to write to some cloud storage like Amazon S3. I'd bet that Amazon's servers can drown JSTOR's, particularly if there's a big pipe like MIT's in-between.

Yeah, it said he went in at least once to swap out the external disk. I assumed it was USB and not eSATA or something.

Also, it's a lot easier to find the owner of a Amazon account (credit card) than a "ghost" laptop under a cardboard box in a closet.

Does anyone think it's odd that an Acer laptop could write these files to disk faster than JSTOR could serve them?

Not at all. Writing to the hard drive is going to take much less time than downloading the articles from a remote server.

Reply on news.ycombinator.com