Sunday, September 8, 2013

Acer Iconia W3, 64GB Windows 8 Pro tablet for $250

Just got this device at Office Depot.

The price was $299, and I got $50 off for opening their credit account. I have to say, for the price, the device is AMAZING. It has a true tablet form factor - 8.1" screen, and not very much frame around the screen.

When I first saw it at the store with the price tag, I thought it was RT, but then I looked at the CPU and it was Intel ATOM, so... it turned out to be the real thing!

Up to now, I did not own a tablet. I don't like severely restricted operating environment (no accessible file system, restricted applications, etc), and this device lifts all these restrictions.

It is still primarily a tablet. You should not buy it as a "lightweight" laptop replacement - it is just not powerful enough for that.

But it's a tablet that can run Office, VideoLAN, and pretty much every utility application written for Windows. It does not restrict your file system access. It can connect to SMB file server in the way normal Windows can - in other words, it is a tablet that does not have all the stupid, artificial restrictions of a typical tablet.

And, of course, it is still a tablet. It runs forever on a battery charge. I don't know if it will last the full advertised 8 hours, but I walked around with it for 2 or 3, and the battery indicator has barely dropped. It has a touch screen, and the screen itself, while not as fancy as on Surface, is very readable.

It is light (MUCH lighter than Surface or iPad), it does not heat at all, and the 64GB version, after all is said and done (and Office installed), leaves you with about 30GB of free space and an expansion microSD card for more. The UI is snappy, and CPU power is enough to play every video that I tried.

And it's just $250. Amazing!

Wednesday, July 17, 2013

An observation for fault tolerant systems

One morning during our vacation stay in California I have awoken to an unpleasant surprise: a front tire went flat overnight. A trivial event, in theory, as the car is fault tolerant when it comes to tires: it has a full size spare.

Unfortunately, the spare was cold: even though the right way to rotate tires is by including the spare in the rotation cycle, in practice I failed to do this, so the spare was simply hanging in its suspension system for several years, untouched and untested.

When I installed it, the very first thing that I discovered was that the pressure in the tire was very low. On this particular car (Toyota Sienna) the spare is well hidden, and it is very easy to forget to check its pressure, so I did. It was not completely flat, but it was not drivable. Luckily, hotel was close to a gas station with an air compressor, but if a tire were to burst far away from the civilization, driving with this spare long distance would have been slow, painful, and unsafe.

Second problem that arose from incorrect rotation schedule was that while the rest of the tires were well worn, this one was completely new. Which means that it was appreciably larger in diameter, making the car asymmetric, and it was on the front wheel. A tire shop could have swapped one of the rear wheels for the flat front, and have the spare installed at the rear, where it would have been less critical, but doing this on a hotel parking lot with one jack was out of question.

By the time the problem was resolved, it took me a good part of the day waiting for Costco to replace four tires instead of enjoying a bike ride across the Golden Gate Bridge with the rest of my family.

So what does all this have to do with the design of fault tolerant systems?

Basically, if the system relies on a cold spare – a replacement part that is squared away, but is not part of the day to day operation, there is a good chance that the spare won’t work – and you will find that out at the worst possible moment, exactly when you need to use it.

Defective spares are not the only source of problems during failure recovery. The recovery process itself is subject to bugs and operator errors. Usually code paths that are activated during recovery are not exercised daily, and can and often do contain bugs that are not ferreted out during regular testing.

If the repair process involves an operator, things can get even worse: an operator also does not execute failure recovery process often enough to be familiar with it, and the probability of a fat-fingered action skyrockets. I personally once lost a whole RAID array at home by replacing a working, instead of the failed, disk.

Most fault tolerance models presume that the failures are independent, and the probability distribution of the second failure is the same as the probability distribution of the first. In practice it is usually not true.

In a fault tolerant system, the mean time to second failure is shorter than the mean time to the first failure.

Since failure recovery adds new code paths and new processes, it is impossible to achieve complete independence of the primary and secondary faults. So… what to do?

A typical reaction would be to ensure that testing failure code paths happens regularly. For instance, the example with the car above had a reasonably simple process-based solution: the pressure should be tested before every long trip. Likewise, testing a master service failure could (and should) be a part of acceptance tests before the release to production.

A better way to handle the situation, however, is by a more careful design.

If at all possible, prefer the design where there is only one role, and if one machine fails, the rest just get higher QPS. This should be the default for services that do not require state preservation, like most frontends. In this case the divergent code path is simply absent, and the code that tests whether to eject a failed system from the query path is always active.

This is not always an option, however. Most backend systems require state persistence, and implement a variation of Paxos, Zookeeper, or simple master-slave protocol where there is a defined leader and one or more followers.

Here a failure triggers a complex leader re-election protocol, and the new leader may exercise different hardware components, which may have already failed, but because the follower did not use them, it is not discovered by the time the election happens.

If the system has distinct roles for primary and secondaries, the simplest way to ensure that all machines can execute all roles is to have it rotate the roles during the normal course of operations. This way a premature switch away from a failed master would be likely to be as uneventful as a routine switch that would have happened just half an hour later.

The leader election protocol would be tested not just a few times in the lab and once in production, but exhaustively, many times under all conditions that arise in real life.

In conclusion: choose a car with a full spare, rotate your tires periodically, and have the spare participate in rotation schedule.

Saturday, January 26, 2013

A big tent of crazy

Today I ran into a page that was talking about what Obama's reelection reveals about America. Which was, of course, the standard "people reject Republican agenda" idea. Nothing new, really, and mostly true, but boring.

What made me chuckle, however, was that while discussing why Obama won, they forgot the reason number zero - that his opponent was a truly terrible candidate. There are probably a few hundred people in the US who could relate to Romney, and they are probably split equally between the two ruling parties. But a few hundred people are not enough to win the vote.

Come to think about it, Romney could have been a Democratic candidate for President just as well - he would have had a similar amount of "base" - which is to say, almost none. He just happened to be a Republican because... well, because that job was available. If he were a Democrat, he would need to switch fewer positions than he had to become a primary-worthy Republican.

The truth is, Obama was a terrible candidate. He promised change, but delivered more of the same. Government secrecy - worse than under Bush. Military budget - bigger than under Bush. Whistleblower prosecutions - more than under anyone ( Income inequality - bigger than under Bush. DOJ pursues a guy who stole a few academic papers, yet lets the banksters who stole billions off the hook.

For a liberal, Obama was uninspiring with a capital U. I could not bring myself to vote for him, and ended up writing in Jon Stewart/Stephen Colbert in his place. Yet not for a second did I doubt that Obama would win. Because even though Obama was uninspiring for the left, Romney was an order of magnitude more uninspiring - for everyone.

I mean, we just came out of a financial train wreck of epic proportions, and here we go, a candidate that runs for a President is a poster child for the forces that created the wreck. What can you expect?

So how did Republicans get a candidate like this?

Because between elections Republicans maintain a big tent. A big tent of crazy. They have people who believe that the Earth is 6000 years old and that the Creation "Museum" is in fact a... museum. They have people who say that evolution and the Big Bang are "lies from the pit of hell" - and probably even more people who do not know what Big Bang is. They have people who think that Obama is a foreign-born Muslim whose Harvard transcripts bear the mark of the Beast. They have people who pray for rain, but don't listen to scientists. They have people who feel safer with the military bigger than the next 10 combined, than with the health insurance. They have people who believe in death committees, want to keep the government out of their Medicare, and want abortion providers to be executed.

After years of gerrymandering, the big crazy camp keeps electing the worst idiots ever in Congress with clockwork predictability. However, once every four years some one has to step out of this asylum to run for the national office - and this is where the process gets confused. People who can win Republican vote have no chance of winning Presidency, and a person who can win Presidency has no chance of winning Republican hearts and minds.

So what to do? So every four years Republican strategists have to source a "more normal" person to appeal to the whole country. After the previous four years in the crazy camp, they only have a very fuzzy idea of what that might mean. Clearly the passion is out - last time a Republican said anything that his party passionately believed in on the national arena this did not end well. Clearly the guy must be rich (because "American Dream", and also because elections are an expensive business). Clearly his personal beliefs must be very flexible - the guy should have a "normal" track record, then win the Republican primary, and then become "normal" again.

And so, ladies and gentlemen, I give you... Mitt Romney. The rest, as they say, is history.