News Stay informed about the latest enterprise technology news and product updates.

Architecting cloud systems and the human side of DevOps deployment

This blog post is part of our Essential Guide: Guide to practicing cloud-native development

According to Adrian Cockcroft, VP of cloud architecture at AWS, one of the biggest challenges in architecting cloud systems lies in architecting the people side of things. Along with the work he does at Amazon, Cockcroft also played a key role in architecting Netflix’s cloud strategy. After Netflix’s rise to prominence, he would often give talks at CIO roundtables who invariably asked where he got all of the superstar engineers. “Well we hired them all from you,” he would respond.

Over the years, Cockcroft pondered what made Netflix so attractive to top talent. One element was Netflix’s very progressive open source program, which allowed developers to take pride by sharing with the larger community of developers. This also runs the danger of attracting recruiters who could lure away programmers.

Compensation and be financial or prestige

Netflix established a very aggressive compensation schedule to address this danger. Each year programmers are rated and compensation is based on what it would cost to hire someone with similar talents. This might not fly in many enterprises with a focus on percentage increases. “I don’t know of any other company that does this on an annual basis,” said Cockcroft.

Another good practice lies in establishing status for important talent by curating a circle of “distinguished engineers” or “fellows.” Once this circle is large enough, members can nominate others into this groups which becomes a sort of elite club. The way to enter this group is to provide value to other distinguished engineers. Over time these distinguished engineers get to know each other as a group, which builds cohesion and loyalty.

Join the conversation


Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

Mr. Hitz, I'm surprised that someone at your level within NetApp has so little grasp on the Isilon feature set including its base clustering capability to seamlessly support 60 second scaling of new nodes, support for heterogeneous node-types, auto-tiering between node-types and it's push-button retirement of node feature - all on-line while auto-balancing for performance and capacity - no downtime required.

Maybe you do understand this but what you're really trying to do is propagate NetApp FUD? Just makes you look a bit foolish.
Wow, knocking EMC for having different feature sets between products. What about NetApp 7-mode, c-mode, and infinite vol mode? Customers not only need to trade off features but also performance to use c-mode and infinite vol. As NetApp tries to catch up with Isilon's scale-out NAS, the feature set of ONTAP diminishes and the performance tax goes up.
Hey Dave, you state you have customers running c-mode, right? Why doesn't NetApp itself run c-mode?

On your June 21st webcast introducing 8.1.1 your own IT guy Randy Stutler stated that "NetApp runs 150 filers in H/A pairs" (that's be 7-mode) and that NetApp IT is just now playing with c-mode in a lab trying to determine "the easy stuff to move first".

NetApp - not a good reference for c-mode. Wonder why . . . .