I’ve been to a lot of technical conferences and this was one of the best. My first trip out of town in my new, used but new to me, car with iPod Pandora wired into an MP3 auxiliary port into the cars speakers (the first car I’ve had with this feature) was a great start and finish. Everything in between was top notch.

I recently had the fortune to peek under the hood at service offerings from Limelight Networks and Oovoo. These companies well represent the continued evolution of digital, networked media and in particular video. Apart from 3D environments, video represents one of the most challenging types of digital, networked media.

Exploding storage requirements are a fact of life. As Moore's law enables higher quality digital multimedia, we need to store more for longer. Keeping multiple systems updated with enough disk storage with RAID redundancy and backup eats up resources. The idea of a SAN (storage area network) is to provide a large pool of storage for many different needs. Economies of scale can be obtained by centralizing fast, redundant storage that can be flexibly parceled out to various systems over the network. Disk storage becomes a cloud resource over the network.

I’ve always been very analytical and grounded in the scientific method: make a hypothesis, set up an experiment to test the hypothesis, measure the results of the experiment, appropriately modify the hypotheses based on the results, and continually repeat. Real world decisions are not that simple. There are many soft factors crucial to decisions in any realistic situation of importance and rarely one simple, right answer. However, good data is an important factor in driving good decisions. Here are 7 guidelines I use:

I recently watched a fascinating presentation about modern CPU architecture: A Crash Course in Modern Hardware. A common theme was addressing issues preventing full utilization of parallel capabilities. My impression from this presentation is that modern hardware speed bottlenecks are often not inherent limitations but rather dependencies between different functions. In other words, one part of the system is waiting for results or input from another, which is waiting for..., which is waiting for...

This reminds me of what sometimes happens in almost any group of people working on a common goal. You often see this on a building or road construction site. The logistics of arranging, sequencing and allocating tasks effectively between multiple people can be difficult and there are also psychological factors.

This past weekend was the opening of a wonderful ballet performance of Alice in Wonderland my wife and several friends were in. It was also the opening of a new server for our TimeTrack media system, the most critical system in our company. As our industry advances and our company grows, we outgrow the hardware our system runs on every few years. Fortunately, Moore’s law makes that pretty easy to address. There are many parallels between opening a show and transitioning a mission critical system to new hardware.

Many people are involved. There is a long period of preparation and practice to make sure everything goes smoothly. There can be chaos, confusion, stress, long hours, late nights and fear as the deadline approaches. There are slips, crisis, and technical glitches during the performance. Breaking down, cleaning up, celebrating, crashing, and a time of withdrawal or post-show blues can follow.

Fortunately in both of these cases, opening night was an exhilarating experience as everything came together and people benefit from all the hard work and preparation. May everyone’s opening be so!

I continue to learn the importance of the user in the success of technology. No matter how good, slick, solid, and scalable a system is, it's worthless if people don't use it. I'm usually so caught up in getting things to work, I underestimate the importance of a simple, clean, intuitive UI, documentation, training and support.

I've always run a Linux development machine in addition to my main Windows workstation. I've had to shuffle hardware for recent projects leaving me with a machine that periodically locks up for my Linux development machine. After weeks of troubleshooting it seems it’s not something easily fixed by swapping out anything except the motherboard.