My Life as a Technology Canary
A gentle nudge from a reader made me realize how long it has been since I’ve posted. Time to get back into the habit.
This has been a particularly busy semester, both from the point of view of my academic life and as CTO. The academic side has been great– I’ve been teaching CS 105, Privacy and Technology, which is always more fun than I should be allowed to have. This is a class that looks at technologies that are seen as privacy-invasive (things like surviellance cameras, wire tapping, and Facebook), dives into the technology and policy, and tries to figure out what can be done. I co-teach with Latanya Sweeney, who really knows this stuff, is a great lecturer, and a better friend. But what made this semester fantastic was the best group of students I’ve ever had in a class–smart, engaged, funny, and fun. On days (and they happen) when I wondered why I was doing all of this, I just had to go to this class to be reminded what fun it is to be at Harvard.
The CTO work has been a lot more scattered, but has also been interesting. Probably the biggest change in my life as I moved to the CTO position was finding that I have very few concentrated, extended periods of time to think about things and get things done. The life of a CTO is one of constantly swapping context, trying to help others (who I hope have concentrated periods of time for their work) to move forward or course correct.
There is also another, odder, part of my job which I characterize as being a technology canary. Canaries were used as early warning systems in mines, organic sensors for dangerous gases. My role of technology canary is to be an early warning system for HUIT on technology trends that are going to change the way we do our jobs. There are lots of these changes coming around, like the changes in client devices (moving from desktops to laptops to tablets and phones, a change that had a pretty disastrous impact on the University of California’s email system). But the most interesting whiff of the future that I’ve seen had to do with a bill from Amazon.
First, some context. All colleges and universities are supposed to offer a net price calculator, a tool that will allow prospective students and their parents to estimate what their college educations will really cost at a particular school (anyone who has to worry about this doesn’t pay list price, at least at Harvard). The financial aid folks here have done a very nice web tool, which they decided to host on Amazon.
Recently, I got a copy of their bill for a month. They had had about 300,000 hits, most from the U.S. but others from all over the world. And the total bill for running this site? $0.63. That’s right, sixty-three cents.
Now, not everything we do at Harvard in IT can be farmed out in this way. This is a simple site, and doesn’t have to be up all the time. It doesn’t have a lot of computation associated with it, and there isn’t a lot of data being moved around. More important, there is no confidential or protected data. But there is a lot of computing at Harvard which has similar characteristics. And at this price, we need to figure out what we can host elsewhere. It may cost more than this example, but even if it is one or two orders of magnitude more it will be less expensive than setting up our own servers and running them here.
This will change a lot of things. We need to figure out what will be changing rather than having it done to us. As the canary, I get to think about these things early on. Which makes life more, um, exciting. But also a lot of fun.
- previous:
- Separated by a Common Language
- next:
- Back to Normal, or Something Like It
thartmann
December 13, 2011 @ 3:29 pm
This really resonates and I can’t agree more, “We need to figure out what will be changing rather than having it done to us” . Far to often it seems as though we position ourselves to react to situations as opposed to being proactive. I would love to see the principles like DevOps embraced in much the same way as ITIL has been for change management. In addition there’s so much amazing talent applying configuration management tools at Harvard, it would be awesome to be able to more easily collaborate and coordinate together, across silo’s and (shudder) use the same code, and even come up with some cross group standards!
In any case, I’m pretty passionate about this particular cool-aid and could type about it for hours, but i’ll spare you! Interesting thoughts!
dsilverman
March 4, 2012 @ 1:49 am
63 cents is a very important take-away, but another equally important is what it takes to get there. Once your AWS environment is setup, provisioning a new machine takes on the order of seconds, rather than the weeks to months that it can take at Harvard. That’s powerful. Paying tens or hundreds of dollars a year per machine instead of thousands would be a huge win, but so would being able to spin up new servers and environments faster, while at the same time cutting administrative overhead (through automatic provisioning, configuration, etc.) and dramatically compressing project time-tables. Plus, when resources are that cheap and ubiquitous, there is less pressure to cut corners in order to save time and money!
Dan Creswell (@dancres)
August 21, 2012 @ 7:37 am
“The CTO work has been a lot more scattered, but has also been interesting. Probably the biggest change in my life as I moved to the CTO position was finding that I have very few concentrated, extended periods of time to think about things and get things done. The life of a CTO is one of constantly swapping context, trying to help others (who I hope have concentrated periods of time for their work) to move forward or course correct.”
I have encountered a similar challenge, I think it’s particularly tough on those with a “from the trenches” knowledge worker history. I understand the value of that time and I’m not comfortable when it’s absent, I feel like my thinking is compromised. That said, I’ve also accrued a huge amount of experience that perhaps means that a lot of the things I used to worry about can be dealt with using some canned behaviour or another. This in turn would mean what little “think time” I get is sufficient in spite of apparently greater load and greater chance of failure. Does that hold up under any reasonable scrutiny? Hell if I know but intuitively it makes some sense to me.
“There is also another, odder, part of my job which I characterize as being a technology canary. Canaries were used as early warning systems in mines, organic sensors for dangerous gases. My role of technology canary is to be an early warning system for HUIT on technology trends that are going to change the way we do our jobs. There are lots of these changes coming around, like the changes in client devices (moving from desktops to laptops to tablets and phones, a change that had a pretty disastrous impact on the University of California’s email system). But the most interesting whiff of the future that I’ve seen had to do with a bill from Amazon.”
Heh, I’d love to understand more as to the oddness you see. I take this part of the role as a great excuse to do technical knowledge work, get back in the trenches a little and stay in touch.