Made a site in college to sell a T-shirt for $1,000,000. Didn't work. But I did make $40 off an MIT professor who bet me I wouldn't try it.
1,344 karma · joined February 20, 2007
Made a site in college to sell a T-shirt for $1,000,000. Didn't work. But I did make $40 off an MIT professor who bet me I wouldn't try it.
But the thing that really irks me is the discussion of caching reverse proxies and partial page caching. Those optimization techniques are language independent. While I came to the same conclusions about usually needing these with Ruby, I fully recognize that it's much slower at running complex application logic and doing in-memory operations than other languages like PHP, Java, and even Python.
Instead of arguing that Ruby isn't slow (because it is), I would have argued that the time to build and debug a Ruby app is less than it would have been using other languages. Extensive library support, an active user community, many great tutorials / code snippets, no need to compile, and a pretty flexible syntax make Ruby great for quick development. But let's not pretend we can have our Ruby cake and eat it too.
haproxy's HTTP feature set is comparatively limited and we simply have too much traffic to put one machine in front of a bunch of web servers.
Take a look at the source sometime. It's kind of fun to see an HTTP parser written like, "if (ngx_str3_cmp(m, 'G', 'E', 'T', ' ')) { r->method = NGX_HTTP_GET;"
Google can justify the expense of their DNS service w/ NXDomain's because it will speed up all of their web properties and generate more search revenue.
We tested it at Justin.tv and managed to cause the main process (gearmand) to spin at 100%, stop accepting connections, and lots of other things. Not quite ready for production use, especially if you need persistence and use the memcached plugin.
Also, the python client library is in pretty rough form. Our developers submitted a patch or two, but it still doesn't reliably handle errors, network outages, failover, or reconnects. You probably want all of these things in your job server of choice.
1) HTTP delivery of live flash content (avoids royalty fees normally paid to Adobe for FMS, so ideally Akamai could lower their prices)
2) Improved analytics (current EdgeControl interface is tricky to use for individual videos)
The rest of the features listed are just current features wrapped up into a nice looking press release. In other words... move along, nothing to see here.
You make a very good point about VBR, thanks for bringing that up!
However, "bandwidth" isn't really what matters. It's the last mile connection to end users. Big CDN's like Akamai have this figured out and can pump millions of simultaneous live streams directly into end user networks without congesting any internet backbones.
It turns out that the main reasons we aren't all watching TV online right now is that big TV networks 1) are afraid of cannibalizing their existing revenue streams, and 2) think it costs a lot money. Neither of these are strictly misconceptions, but they aren't nearly as true as they were even a year ago.
Justin.tv has developed technology and infrastructure that brings the cost of distributing live video below the cost of serving static video, so I guess that makes it all about the revenue streams. It should be interesting to see what happens to this space in the next couple of years.
You're making the right choice on 1U 8-core boxes. We've done the math and there's no better combination of price, performance, density, and power consumption out there.
-Kyle
Classic.
However... we have our web servers in a colo. Ping times to EC2 aren't fantastic, so we use a traditional CDN for static content and a colo for application servers in order to keep things snappy.
I had a development setup on the bench that was running on a mini-itx board. PG convinced us we needed to launch the second my prototype worked, so I literally designed a case around the bench setup and stuffed it into a backpack.
Justin now carries a 2 lb Sony VAIO that runs Windows XP. The technology is simple enough that anyone can use it, and soon they'll be able to.