space-monkey

"That's just semantics."

Exactly. The semantics is what the words actually mean.

Most people don't upgrade their CPUs. Most people don't know what generation their CPU is from when they buy.

The tradeoff is in that AZs are "engineered to be insulated" as opposed to being actually, or naturally isolated. Prior to their downtime, I've had plenty of conversations with folks that I work with about AWS and we've always assumed that AZs are not 100% isolated. I can see how someone can read "engineered to be insulated" the other way, but I generally read these kinds of materials as guaranteeing nothing beyond the most limited possible reading, and probably not even that.

The quoted statement doesn't say that isolation is 100% or that multiple AZs can't ever ever fail at the same time. It says that if only one AZ goes down and you have servers in another, then those servers will still be up, which should be obvious. Insulated doesn't even mean the same thing as isolated.

They don't have to take it.

And that's why they have multiple fully-isolated regions. Availability zones are a purposeful tradeoff that provides easier to use service with higher inter-zone communication performance and lower cost.

I'd go see that movie.

It's not a huge restriction since GAE will happily spin up dozens of instances without any intervention if your load requires it. Getting multiple concurrent requests into each instance will make things like local memory caching somewhat more useful though.

There seems to be a steady stream of these from ex-Googlers, which is frankly what I'd expect from a company of that size. What I'd be interested to know is what percentage of SWE's are leaving each year, and of course how do the people that are left feel about the company. The ones that I've heard from have a range of opinions from It's Ok to It's Great, but those are of course anecdotes.

Also, "behaviors that lead to success" are not necessarily the same as "behaviors of the successful". Lots of folks have effectively won the lottery.

Programming is not tedious, and it is hard. Listing instructions for the computer to follow is only 10% of the job. Most of the job is figuring out how to deal with complexity in a way that is economical, maintainable, and satisfies secondary requirements like security and performance.

In any system design involving software, the software is where the complexity goes. To save money, you might replace an electronic component with an algorithm. To make a better product you might replace a tedious user action with an automated system. This is because (some!) software developers/programmers have tools for dealing with complexity. Things like abstraction and re-use that don't trivially map to the concrete world.

Frankly if software were about enumerating steps, I would have been out of this business long long ago.

Not if x is negative.

Reflector is a great tool. They should charge if they can.

You can't continue to get affordable insurance if you are continually making claims.

I bought a Mac Mini and returned it last year. If it had cost 1/2 as much, I wouldn't have returned it. It just didn't do enough for me to justify the price. If something turns out to be worth less to you than (the return value - time and irritation cost of doing the return), you return it.

Last time I heard, twitter still uses MySQL for the statuses (tweets) table. They did have a plan to migrate to Cassandra, but didn't go all the way through with it. So I find it hard to agree with "you just can't scale those". It may be a lot of work, but for some applications you can scale them.

Edit: twitter cassandra link (don't know if this is the latest): http://engineering.twitter.com/2010/07/cassandra-at-twitter-...

Agreed. These are good places to start because you can just start building and learn as you go. Once you start building, you'll figure out what you need to be reading very quickly.

And IIRC Oracle had been actively attempting to exert this degree of control, to the point of interfering in infrastructure changes that project contributors wanted to make.

I believe you can enable billing and leave your cap at $0.

Frankly, the great thing about leading edge datastore development right now is that it's all over the map. Many of them will fall off the map over time, but having a couple of additional models to add on to the current great RDBMS tools can only be a plus.

I imagine that a significant driver of datastore heterogeneity is that there are a number of very different in-house datastores that are supporting extremely successful commercial ventures. It used to be that "you don't get fired for choosing Oracle." Now that Amazon and Google, etc, have paved the way, it's much more feasible for a corporate infrastructure developer to try to find a way off of the high-dollar proprietary systems like Teradata.

I also think it's interesting to note that some (most?) of the largest scale commodity RDBMS users couldn't get close to where they are without something like memcached.

It's not legally harassment, but it's definitely not calm, rational, productive communication either.

And the 32-bit address space generally has some space reserved for the kernel, so more like 3GB per process, depending on the OS.

Be unfortunate enough to get seriously ill without medical insurance.

I see over and over again that "Relational databases are not easily scalable. It will require a lot of efforts to build a scalable and redundant setup." I would like to see more mention of how many users/rps/etc you can handle with a maximally sized Amazon RDS redundant instance.

Scalable is different than infinitely scalable. Scalable used to mean, for example, being able to get a close-to-linear speedup from multiple cores, more or faster disks, etc.

I'm curious what the latency impact is, and how that compares to more traditional high-speed routers.

If it were "ideal", it wouldn't be game-able. I'm not going to claim that this ideal is possible!

"if Heroku simply spawned a new instance (they're already running on EC2) for each customer, wouldn't this be mitigated?"

Yes, but that'd cost a ton more than what they're doing right now. Consider the number of apps that they are hosting at the free level right now. Offering private instances as an upsell might make some folks happy though.

Google is probably right to not go down the hard-coding path. Hard coding the offenders doesn't work. You block a domain, they changes names. You block an IP, they switch to AWS so that your choice is to block all of AWS or not block them.

Agreed. My ideal search engine wouldn't require real websites to play in the SEO arms-race to beat out the junk sites.

"You'll never go back to school - The statistics don't lie on this one: people who take terms off are disproportionately unlikely to finish up, even if its only a couple classes."

Is this just evidence that once people start working, they find the value of the degree to be less than cost of finishing it?

You could let the user designate a charity and get friends to pledge donations if the user meets their goal. A smaller charitable gift may go a longer way psychologically for some folks, and it may be easier for them to get their friends to pledge as well.

Then, when you go to eat the bowl of ice-cream or skip your workout, you think: do I really want to screw the panda bears/starving children/cancer sufferers?