Last week, Apple held WWDC 2014, their annual developer’s conference. A number of tech analysts have been bemoaning the lack of new laptops, beating the “Apple is doomed” drum because Apple didn’t introduce a supersized iPhone, or losing their minds and credibility by saying that if Apple doesn’t release a watch/fitness device in the next 60 days they will become irrelevant.

Since Apple dropped attendance at MacWorld, there are now pretty much two Apple events every year that you can count on: the iPhone Event in September and WWDC in the summer. There are often other events sprinkled in, depending on product readiness and release cycle, but those are the two that you can take to the bank. The a lot of industry analysts have come to expect hardware miracles every time an iPhone is released, and now, every time WWDC happens. They are used to Apple holding it’s cards very close to it’s chest.

What they missed is that Apple has played it’s hand. In full. What they were announcing last week was, as best as I can tell, a New Apple. This is the point where people will look back and say “2014 was the year that Tim Cook’s leadership really started being evident outside of Apple”. Not in the supply chain side of things, but in terms of making Apple focus on tearing down silos and building a better sense of community and collaboration, both within Apple and with the larger developer community.

Apple has given developers a lot of the things they have been asking for for years. Opening up inter-app communication on the iPhone is the most obvious part of this, and everything it enables. App store metrics, beta testing services, new languages, closer integration between the desktop and mobile, access to the raw data in iCloud and so much more. Pretty much they only thing they didn’t do was replace HFS+ and provide an SDK to program for the Apple TV.

Apple is also taking a more realistic approach to it’s fabled secrecy. The NDA’s on the conference, for example, don’t prohibit attendees from talking about the iOS 8 or Yosemite betas. All the conference videos are posted – along with their slide decks – out in public. You don’t need an Apple Developer login to get to them. You don’t need to sign a NDA to agree not talk about the session content.

That’s not to say that Apple isn’t being secretive. I’m not sure how many companies can keep a whole new language a total secret for as long as Apple did with Swift. Nobody really knows what they are doing with sapphire crystal. The Apple TV and the Mac Mini are in desperate need of a refresh, but there’s nothing leaking about where those products are going.

Personally, I’m happy Apple focused on the Developer part of the World Wide Developers Conference. That’s where the focus should be. We’ll see where all of this goes, and if it’s part of a larger trend or just a one-time event. I’m betting that this is a New Apple, and that we are going to see a lot more WWDC’s in the future like this one.

I don’t see much of a need for Dropbox anymore. It costs a lot more than the alternatives ($0.10/GB/month vs $0.02/GB/month), now that Apple has announced prices for iCloud Storage that are on par with the Google Drive storage prices. Further, the Dropbox iOS app doesn’t sync for photo bursts on iOS7 (we don’t know about iOS8 yet), and Dropbox seems dedicated to bringing out new services like Carousel and Mailbox – which are services I don’t want. Finally, if Apple actually fixes Photo Stream, I won’t be using any of the features Dropbox provides other than occasionally sharing folders with other Dropbox users. I’m pretty confident that my shared folder use will fit within my free Dropbox quota.

To further compound issues, if Photo Stream actually works in iOS8, I will probably move my file sharing over to a self-hosted copy of ownCloud. No storage limits, and the files are local to my fileserver. Which can easily be backed up to CrashPlan. There is also a chance that I will be able to use iCloud Family Sharing to share purchased quota, even further driving down my monthly costs.

Of course, all of these are forward looking statements based on a product announcement yesterday. No changes will be made until after iOS8 ships and I can see if it works in practice.

Time to report another website credential compromise – this week’s winner is eBay. There is a nice writeup by Ars Technica. Go change passwords, folks. And if you don’t already, look into something like LastPass or 1Password.

I use the BeagleBone Black as a low power Linux server that can control temperature sensors and do other systems administration tasks for me. One of the issues the board has is there is no on-board battery, which means the system clock resets every time it reboots. I ran across an article that suggests that the BeagleBone Black has headers and the proper circuitry to charge batteries.

As I move forward with oVirt, I’m going to need ways to manage the power to the oVirt nodes, as I’m planning on building a case and not having traditional things like ATX power supplies or power switches. A BeagleBone with a power relay and a battery that can keep it alive for a long time works for me. Time to make another Adafruit order and get comfortable with soldering very small components.

It also means that I can move forward with a very old project of mine, of having my bike track it’s location and the air temperature with a BeagleBone. I wasn’t happy with the complexity and cost of wiring up a battery charging circuit to my Pelican Panniers, but this looks like a win.

UPDATE: (06/16/2014)

This project is on hold until after I refresh my fileserver and move the environment to our new house. There are significant parts of the storage stack that aren’t cost effective enough at the moment to allow me to get 12TB of usable space online, have multi-node redundancy and bring up a multi-node virtualization environment at the moment. For now, I’m just going to execute on my FreeNAS upgrade and noodle around with a couple of low-power embedded systems. I will likely return to this project later this year or early next spring.

END UPDATE

As a preview of the tech projects I referenced in my last post, I’m setting up an oVirt cluster on my home network to help me better understand how oVirt works. The idea is that it will let me manage the various resources provided on the home network in a sensible fashion. This includes file services, backup services, and Plex when listing the visible stuff, but also includes DHCP, DNS, temperature sensors, external web services and ownCloud.

There is also the desire to really get down to the meat of this all – set DHCP options, have machines PXE boot to gPXE/iPXE and then come up off an iSCSI volume provided by my fileserver. And that’s just one part of it. I know I can do it – I’ve used these systems at work to death – but I’ve never owned the whole stack myself.

The primary enabler/driver of this is AMD’s new AM1 platform – they are cheap enough (and powerful enough) to be nodes in an oVirt cluster without breaking the bank. The motherboards are $30-$50 and the CPU’s are $30-$65. This means that with a cheap case, a power supply, CPU, RAM and motherboard you are looking at a total cost of $210 or so for a diskless node. With the APU’s TDP of something in the order of 25W, a silent 12V DC-DC ATX power supply starts to look incredibly tempting, especially if you aren’t spinning platters. If you forego the case and like to make modifications to IKEA furniture, you can do something like the Helmer cluster in a file cabinet – which is almost exactly what I want to do. Not getting a case or an AC-DC power supply drops the per node price to $195 or so, each. This is for 2Ghz x 4 core, 8GB node with a TDP of less than 35W at full draw.

The things that need to be figured out, before funding any of it, is what I’m going to need in terms of I/O requirements from the fileserver. The plan is to move to RAIDZ2 for the main file store, and set up the old 750GB drives in a triple parity mirror to maximize random read IOPS, which is the primary need for the VMs that will be consuming that pool. I’d like to get away without having to budget in SSDs for L2ARC, but that means buying pricier 8GB DIMMS.

That’s the direction I’m moving. As the project moves along, I’ll post. There are other projects as well, but this is the one that is the furthest along.

I’m starting to gear up for some personal tech projects – there’s a number of itches I want to scratch, and the itch is getting stronger. Part of the motivation was running across the “12 factor app” website a few days ago, and thinking about how to design modern IT and web services. Anyone who works in IT doing app or website development, please take a look. Much like RESTful API development, this isn’t a how-to, but more of a cultural framework of why you should build you apps in specific ways. A lot of it has to do with preparing yourself to scale out instead of up, and help you retain service and platform portability.

Today is the ninth anniversary of my photo-of-the-day photoblog. It’s not always been a photo of the day (I skipped Saturdays and Sundays for the first few years), and I took some breaks now and again. It’s always fun for me to look back and see how much I’ve improved over the years.