...but then again, I couldn't have topped Microsoft anyway.

I've joined Microsoft's Tech Preview of Live Mesh, and although it's very early in the product's life cycle, I'm already getting value from the tool. One of the ideas that a friend and I bantered around last summer was to solve the problem that file synchronization solutions like Groove have in dealing with devices being off line. We wanted to create a synchronization service that would handle the files "in the cloud". The only difference was that we wanted to only store files until they either expired or were pulled down to the devices that were not connected at the time of the original upload. We were being pessimistic about our ability to manage the sale of storage required to support a fully accessible file store for the web.

Regardless of the argument about Microsoft's intentions (does the Mesh support the "cloud" model, or defend against it), they are going to have a hit. To me, the "cloud" model has it's limitations and the local desktop / OS has other problems. Merging the two is what will get people into the cloud way of thinking.

Labels: cloud computing, live mesh, microsoft, web 2.0

I lost access to my work email account yesterday afternoon, and it has yet to be restored (IT has been working on it since then, but they haven't figured it out yet). I always laughed at the idea of email addiction, but it really is painful to try to operate in a modern business environment without that trusty "send" button. I'm up to four calls today where people have assumed that I have already read something that they sent me. When they say this, I have to sheepishly reply that "I can't get into my email today, so I haven't seen that yet." When I go on vacation, I happily ignore all of my email accounts, but this is like walking around a cave without a light.

Labels: personal

There's been a ton of commotion recently with everyone getting excited about Google's App Engine and Amazon's EC2. Rackspace has been getting into the mix as well with Mosso. It's pretty exciting stuff, and has the potential to take a significant market share from the more traditional hosting models (shared, dedicated, managed, etc...).

Lew Moorman (from Rackspace) argues that there will always be a place for the other models. I actually think that he's right in the short to medium term, but fundamentally, computing power has become a commodity. It's getting to the point where individuals and forward thinking organizations are abstracting the details of hosting their applications away from themselves, in effect reducing the level of system administration knowledge that they will have to keep on staff. Like it or not, that's not going to stop.

Programming language evolution is a perfect example of how this abstraction moves certain technical knowledge domains out of the mainstream and back into being a speciality. How many programmers do you know that actually use any sort of assembly language programming in their work today? I'll assume not many. Instead, programming has been abstracted away from the hardware and into the domain of solving user problems.

Looking at the hosting market as it stands today, we are going to see some serious change over the next decade (if not sooner). Today, many companies are outsourcing the server and network administration work to an ever decreasing number of service providers (isn't consolidation fun?). This is the same principle as the programming example: companies are abstracting the detailed infrastructure work away from their staff. But in many ways, it isn't really abstract as much as the executive suite would like to think. Most enterprises still outsource in "server units" or "network device units", ensuring that they have to think in terms of physical or logical servers.

Now look at "cloud hosting". It's another abstraction layer for application infrastructure, outsourcing most of the challenges of the physical systems completely to the providers.

So here's where it's going... if cloud hosting providers can solve a couple of key problems, then they are in a position to help enterprise buyers move to this new model. Issue number 1? Security. In this world of industry competition, government regulation and consumer data protection concerns, this is of paramount concern to enterprise buyers - which is why it should matter to cloud hosting companies. These hosting companies should start to have a dialog with their various technology partners and potential customers about the following topics: Data Privacy, Authentication, Authorization, Logical "Tier-ed" Separation of Application Layers, Private Network Transport. The sooner this conversation starts, the sooner the problems can be solved.

NOTE: This posting talks about the hosting business... which my employer is actively involved in. The ideas expressed above are my own, and have absolutely nothing to do with the company I work for.

Labels: cloud computing, hosting, virtualization

Apple is apparently picking a fight with New York City's GreeNYC campaign over logos (see the Wired article). This is just silly. I think that apples have been represented in art before.

Do these logos look the same to you?

Looking around for a good way to illustrate our team's branch / tag model more clearly, I ran across a recent article by Henrik Kniberg titled Agile version control with multiple teams (PDF) (he also published it as HTML on InfoQ here).

Henrik has done an outstanding job of describing and visualizing the thought process around concurrent team development within the same repository. I'm particularly impressed with his simple explanation on how to think about your release mainline's relationship with multiple development code lines.

Labels: configuration management, process, project automation, release

Recent events reminded some of us about the IBM commercial on "ideating". This cracks me up every time I watch it.

One of Google's April Fools jokes for 2008 is the Google Wake Up Kit. It's definately not as funny as Google Paper.

For something actually useful, check out the Google Docs "Offline" capabilities.

Labels: Google