I'm working on a way to create a better understanding of DevOps for our projects, as well as some basic primer readings. In the meantime, here is are two keynote talks at LISA conferences. The talks are long, but well worth the time. is over an hour, but 10-12 minutes in and you'll get the main idea. Stay with it as long as you can. Would love to hear your points of view on this.
- Here is a short video summary of the history of DevOps began. I learned a lot. http://itrevolution.com/the-history-of-devops/
- Gene Kim gave a keynote at LISA 2014 (Usenix special interest group for system administrators - mostly ops people). This is considered a seminal talk. https://www.usenix.org/conference/lisa14/conference-program/presentation/kim. (Some of what Nicole Forsgren talked about in my class is mentioned this talk.)
- Also Gene Kim et al., wrote "The Phoenix Project" - read Inside Amazon. Worthwhile and easy to digest.
- A really fantastic talk was given by Jez Humble at LISA 2015 (keynote), delivered to an Ops crowd. He talks about how DevOps is different from Agile Lean Configuration Management
The talk also speaks to project management and how faulty the thinking is for understanding systems. Jez defines IT high performance. He gets to this very early in the video.
ReplyDeleteIt is evident that the unfortunate concept of "if you want to go faster you are going to reduce quality" applies in many different industries. Not every industry or company has contemplated going lean. Car manufacturers, food suppliers, and even small locally owned restaurants struggle with the want to
ReplyDeleteget stuff done faster, keep quality high and reduce the price all at the same time. Changing the process in which processes are completed, allotting more time in the beginning of a project for testing and debugging, and communicating across departments could change this problem as described in this video.
Good points. Consider, too, that it's possible to produce great "quality" from a software developer point of view (e.g., fast, tested, stand-ups, etc), yet it doesn't consider the operations "over the wall", so fails in the long run. Great examples about food suppliers and cars. Most industries. Examples in car industry are failed air bags, brakes, multiple recalls. Cars (like software) rely on third party vendors. Testing and debugging should be done on the floor, but time is money. Now cars will rely on computer programs, so the consumer might be the one who determines if it "works."
DeleteI happen to think it is more fundamental than testing and debugging, but communicating and working with those who will be affected by the change - i.e., those who have to implement the well-tested, debugged work. My 2 cents.
I posted this on our course readings (and schedule), but linking to it here. This is a DevOps case about Standard Bank.
ReplyDeleteLooking at the video on DevOps I realize that our MIS curriculum perfectly supports the approach of DevOps. With DevOps revolving around the idea that both development and operations need to be brought together to collaborate in project management, the MIS professional has the perfect skillset to bridge the gap between these worlds. I consider the MIS professional to be with one foot in business (operations) and with the other one in development.
ReplyDeleteAn integrated approach is necessary as I still see many companies approaching projects from just one of these worlds. Often I see job descriptions for developer related aspects with the need to posses special operational degrees. I've also seen the opposite where developers design a system without consulting operations. I think these worlds need to work together preferably with someone in between, like the MIS professional.