Link dump

[evwhore]

Heh, for my vegetarian friends, response #2.

via boingboing, Out of Context Science -- sentences taken from scientific papers that are funny out of context. LJ syndication: outofcontextsci

Remember this guy? He sued the DEA, saying that they leaked the video, violating his privacy because "the video's distribution resulted in him becoming the 'target of jokes

Comments

[clutch_c]

"Got memory? Oversupply Sends DRAM Prices To One-year Low"

I was amused by those prices ($0.81 for 1gb chips). I remember in 1981 when 64K chips were introduced to replace 16K chips at $20 or so each. Costs have decreased since then by a factor of 400,000 (ie 1gb of 64K chips would have cost $300,000+).

He said, apropos of nothing.

[evwhore]

I remember in the mid 80s buying individual chips that I then had to insert into slots on an expansion board myself for an Apple //e.

I am semi-regularly amused when our monitoring processes notify us of "low disk space" where the amount of free space remaining is orders of magnitude larger than the first hard drive I bought.

[jnala]

I very much doubt that the Akamai story is accurate as written. An alternate explanation is that the poster has an ISP with shitty peering, and using a non-local DNS resolver prevents Akamai from sending him to a local server and saving him from his shitty ISP. This seems far more likely to me than hypothesizing that Akamai is a bunch of idiots who both are unaware of the existence of public DNS resolvers, and have no clue how to do load balancing despite that being at the core of their business.

[evwhore]

I'm not sure I quite follow. Am I rephrasing your scenario correctly:

1. use ISP's DNS, Akamai sends you to server local to your ISP
2. use public DNS, Akamai sends you to a random server anywhere in the US?

and that because the poster (hypothetically) has an ISP with shitty peering then (1) is reasonably fast because it's local but (2) goes further upstream and is therefore slower?

I suppose I follow that, but really? 360x slower?

EDIT: I looked at the comments: "First, I was able to verify this with the iTunes download. My Cox DNS was 20 seconds while my Google DNS was 2 minutes 10 seconds." OTOH he had the opposite experience with an Adobe download.

EDIT again: a lot of back and forth at the /.-linked article. I find it plausible that it's a combination of factors and not simply IP addr of DNS client.

[jnala]

Eh, there happens to be network congestion or a routing issue somewhere between poster and random server, it could easily be massively slower.

If the "everyone using Google DNS gets sent to the same Akamai server" hypothesis were true, then everyone using Google DNS would have the same absurdly horrible performance, and it doesn't sound like that's what's happening, from the random snippets of anecdote we're getting on discussion boards.

Assuming that the requesting DNS server is local to the client and handing out an address local to it is a hack, and obviously it's not what you'd do if you were building a CDN-aware network architecture from the ground up, but in most cases it's a hack that works pretty well. You devolve to the pick-a-random-server case with public DNS users, and you get support calls from people every so often going "CNN.com is slow as hell! yeah, my DNS server is in Switzerland, so?", but for the most part, for most users, it works.

(I don't remember if you know that I used to work for Akamai, btw.)

[evwhore]

I think I might have known that but if so I'd forgotten it until you mentioned it just now.

iOS 4

[ultrapower]

The fucking iOS 4 alarm bug affected me. I set an alarm for 2011-01-02T10:00, and it didn't ring. At 11:30, I assumed I had just slept through it. Apparently not.

Well, technically, I did sleep through the UTTER SILENCE, but you know that's not what I meant.