Alex Hopmann’s Blog

I’ve been having a problem lately with the Windows Media Center I use for my main TV. I’m still running Media Center on Vista and I have a 750GB drive in the box that only has about 200GB of real content on it, but even though it should have around 500GB free, its only reporting 10-20GB free.

My suspicion is that the Windows Home Server integration might be at fault. I recently upgraded my Home Server with Power Pack 3 and it has a cool new feature that integrates with Media Center and can automatically archive data to the larger disks on the server. The Home Server has always been a bit strange about how much free space it reports- despite ~7TB of real free space, the system disk I have is fairly small and I think that causes it to report only a small amount of free space. Now that the C: drive on my Media Center is doing the same thing it makes me suspect that the Home Server is doing something fancy with the disks and causing this problem.

Anyone have any ideas? The Home Server tie might be a total red herring…

As I typically do this time of year its is time for “Winter PC Cleaning”. This is the time of year when I often end up looking at which machines around the house I need to upgrade, which to replace with new machines, and which machines to reinstall Windows to get rid of the cruft that tends to build up over time.

On that last point, its unfortunate that its necessary but it really does seem like a good thing to do. It isn’t as bad as it was in the old days when your system really got slower and slower and progressively less stable if you didn’t clean it up. But as improved as it has been the last couple of years, there is still plenty stuff you install that makes various hooks into the Windows Explorer, into your codec list, and all over the place and without starting from scratch now and then. In the past I’ve felt like I had do to this every year with my main machines, but I guess its a mark of progress that I’ve survived for a whole 2 years without worrying about it.

Two years ago I built my new workstation using top of the line parts of the day. At the time a 3ghz (overclocked to 3.66-4ghz) quad-core CPU with 4GB RAM was pretty much top of the line (unless you went to the server-class Xeon chips in which case you could use two quad-core CPUs for 8…) It is an interesting benchmark of the state of progress in the industry that while the shiny new Core i7 machines are indeed faster, they aren’t enough so that I’m really eager to build out a whole new system right away.

Still there have been a couple of things that have been annoying me. My home workstation was still running Windows Vista and while I’ve been increasingly using a Win7 VM on it, having used Windows 7 at work for a long time now I’ve been eager to upgrade it. The other biggest concerns were that the main disk has been feeling really slow. Its one of the expensive faster “Raptor” drives, but its not that much faster and combined with the fact that I’ve been running 32-bit Vista which sure seems to page a ton, and that the drive is so noisy you really hear it when its seeking all over the place, it has felt like a real dog.

So given that I’m moving up to Windows 7, I bought myself two key upgrades- more RAM to take the system to 8GB (which with the 64-bit install of Windows 7 I plan should work nicely) and a SSD for my boot drive. NewEgg was nice enough to have a screaming black-friday deal on the 80GB X25Mg2 drive for $219 (down from the usual $279 or so) which was enough for me to pick one of those up right away. I kind of wish I’d bought two so I had one for my laptop, but I’m not quite going there yet.

The other big plan is to install Windows 7 in a VHD so that I’ve got more flexibility to backup my OS image. I found some instructions here although I’m not worried about the dual-boot thing. I know some friends who are using dual-boot to have separate partitions for non-released stuff (beta versions of Office and other stuff) but I just run those in VMs and don’t mess with different base OS images.

There was one big trick missing from the instructions for VHD install. When I went to setup Windows and selected the VHD disk it told me that the given partition was not supported for boot. Turns out that the “Next” button was enabled after all and the install worked just fine, so that warning just needed to be ignored.

So hopefully that should do it for now. I suspect it will be time to upgrade my video card (currently an ATI 3870) in a few months, although I’m really hoping that they will come out with some that support 4-6 outputs on a single card (which is the rumor). Those of course might need new connectors (DisplayPort) so the scary part might be waiting on a new generation of displays. Since large multi-touch displays should also become practical sometime in the next year, I’ll be waiting on those for sure before I upgrade anything.

One interesting twist- when I was doing my table I was mostly thinking of SATA in terms of the high capacity 3.5″ drives, where you can get a 1TB drive for ~$90 and a 1.5TB drive for $120. So that is actually $.08/GB (which I rounded up to $.1). But a very interesting in-between solution is to go with the commodity 7200RPM 2.5″ drives. The fancy 10k/15k RPM SAS drives are always worse performance/$, but a $60 250GB 7200 RPM 2.5″ drive gives you pretty much the same IOPS. Plus I’ve seen 2U server designs that can fit more than 25 in a single box (as compared to ~8-10 max 3.5″ drives).

So if you aren’t trying to maximize capacity in a 2U unit you can either do-
10×1.5TB 3.5″ drives, 15TB storage, 750 IOPS, $1200, 0.625 IOPS/$
or
25×250GB 2.5″ drives, 6.2TB storage, 1875 IOPS, $1500 1.25 IOPS/$

Which represents twice the performance/$ at a cost of less than half the capacity/$. But again, if you weren’t going to be able to use that capacity anyway, its a great trade-off.

Alternatively in the same price range you could get two of those SSDs if you data-set is really small-
2×64GB SSDs, .13TB storage, 6600 IOPS, $1400 4.7 IOPS/$

So the SSD performance/$ is still better, but the capacity is so small that its unlikely to work for many applications.

I was doing some analysis recently and I put together a simple comparison chart between commodity SATA drives, enterprise SAS drives and enterprise SSD. Often SSD is positioned as “fast but very expensive”. But expensive is all a matter of what you are measuring. SSD is VERY expensive for capacity, but realtively inexpensive for performance.

First numbers in each cell are normalized to the SATA drives.

The table above makes it pretty clear just how different these technologies are. The SSD is 100x more expensive per GB than the commodity SATA drive. But when you measure based on IOPS it is 6x cheaper. You would need 44 spindles of traditional drives to match the I/O performance of a single SSD. Before I go on, I should point out the measurements aren’t apples to apples. The SSD measurement is write performance. The SATA performance varies a ton depending on the type of operations. Sequential reads and writes can be quite efficient while a SSD doesn’t benefit from larger reads & writes beyond a certain point.

It is also interesting to observe that the SAS drives appear to never make sense. They are 50% faster than the SATA drives but they cost so much more that it is hard to imagine a scenario where they would be the right choice.

For someone with such crappy handwriting its surprising that I’ve always been so interested in using it as an interface to computers. Most of the time I don’t care at all about voice user interfaces- it seems like a giant pain to talk to my computer (although some of the phone scenarios are very interesting). But let me put it this way- I have one of the Newtons from the first day the first model went on sale.

Of course the reality of handwriting interfaces to computers has been pretty much a continual disappointment. I was hopeful about the whole idea of the TabletPC since it seemed like a much better way to both have a device for reading plus have a computer in meetings without the effect of everyone sitting around in a room with a screen up in front of their face. Somehow having a tablet sitting on the table feels less rude to me (although I suppose it can easily be just as bad).

Unfortunatelly the form factors of early Tablet PCs and the user experience of the handwriting input has been bad enough so far that it hasn’t been worth using. Ever the hopeful type when I needed a new laptop at work I opted for a Tablet and got the HP 2730. Now, the 2730 isn’t great- I’ve got a long list of complaints and let’s just say that it isn’t really competing with a Kindle as a reading device.

But the great news is I’ve been running Win7 lately and I’ll say that the most noticeable breakthrough in Win7 for me has been how well the handwriting user experience works. It still isn’t perfect, but the current flaws are more like minor oversights than huge flaws. For example, when you go to click on the button in the top left of Office apps, the handwriting panel is right there and likes to pop out. But overall the experience is 1000% better and so far has finally gotten good enough to be solidly useful. I wouldn’t want to write a long document using the handwriting, but mostly because I can type about 5x faster. But for responding to brief emails and IM messages, and general interaction with the system it works great, the handwriting user experience seems smooth (I even like the font they use as it translates your handwriting) and so far I’m still using it and expect to keep using it.

It will be interesting to see how things work out with the new in-app purchases and all. I think there is a certain amount of fear that the whole app market is going to turn into a bunch of semi-scams where every app author tries to constantly squeeze more money out of people. Part of the biggest danger is that app updates which I tend to just do automatically can regress functionality.

One of my favorite casual iPhone games is FlightControl. I think it cost $0.99 and its simple, easy to pick up for a few minutes while waiting for a bus or something. I noticed a week or two ago that there was multiple levels in the game- you could switch the map and try out other environments. Now that appears to be gone- probably something that looked like a routine update removed it and I fear its going to be back later as an extra charge…

I got a new iPhone 3GS today (actually Kat got it, but is letting me have the new one while she gets my old one that isn’t eligable for an upgrade yet). They were warning that the activation is taking forever but after we were home and an hour later it still wasn’t activated.

The LA Times blog had some advice to turn the phone off and back on. I gave it a try and 2 minutes later it activated right away.

I’ve got to say, this year’s iPhone release madness was much better than the past. The lame upgrade policies (even the recently ammended ones) suck- I do expect that for my $1000 a year in service + $299 a year I should be able to get a new phone every year. But the reservation process was great and even at 3:30pm I was able to walk into the store, after a 2 minute wait someone took me in, they grabbed the box and it was all very smooth. I’m especially impressed that even on the most busy day of the year they still had things worked out enough that someone was able to notice us puzzling over the display of protective cases, approach us and helpfully suggest a few models, unbox them for us to try out and all. The Apple Store continues to the the retail gold-standard.

Uploading any files more than a few hundred K from a browser has been a problem for years. The UI available in the browser is very limited and relying on a single HTTP request that might take minutes or hours (and that you have to start over from scratch if it fails) often turns into a huge source of user frustration. There is also an extra flaw in that the TCP connection can fail before the whole file is transmitted but depending on the circumstances the server might not be able to tell if the whole file was actually received. There are a bunch of sites that use various ActiveX or Java controls but those have typically been a pain to install and/or flakey.

I just discovered that Silverlight can be used to create much more functional upload controls. Here is one for example in the Codeplex Code library. Granted, users need to have Silverlight already installed, but once they do it becomes much easier to have a good user interface, while having the actual process send chunks of the file that can be resumed if any piece fails, etc. Combine that with the Azure Blob chunked-PUT mechanism and you can build a very robust storage mechanism right in the browser. I’m looking forward to trying it out.

One last thought- it would be useful to define a standard protocol for uploading content in chunks (and yes, this is distinct from an HTTP PUT/POST with chunked encoding). Something along the lines of what the Azure Blob store does but defined as a standard that various controls and services can all interoperate.

Coding Horror has a post highly recommending the new book “Regular Expressions Cookbook”. Now, I have mixed feelings about regular expressions and get concerned when I see them since they are often overused and when misused can result in code that is very hard to understand and debug. Having said that, when used in the right situation they can be a perfect solution to otherwise complicated text parsing & validation. But the art of creating them is often a lot of voodoo, so a book that has good reference materials and examples would be very helpful. I’ll report on this one next week after I have a chance to look through it a bit.

As an old HTTP guy I often get nervous about new services. They tend to violate all sorts of key HTTP architecture concepts and just take advantage of the flexibility to do whatever.

I’ve been really happy to see that the Azure Blob storage actually gets this stuff right. They have valid REST semantics with a good URL namespace, support GET and PUT with the right kind of range headers, etag and conditional operation support. They have a smart design for uploading a large blob in multiple pieces (which works around one of the bigger flaws in the older WebDAV support), and all. Anyway, its great to see a team do all their homework and get these details right- I suspect this will really payoff over the long lifespan of a service as it fits in cleanly with rest of the web services world. (note- I’m not saying other competing products aren’t also doing these things right, I haven’t researched those details lately).