Devops is related to bringing Development and Operations closer together.
Tuesday, October 25, 2011
Kanban Weekly Roundup - Oct 25, 2011
By Dominica DeGrandis
After attending Devops, Lean and Kanban conferences in Sweden and Germany, this week’s roundup is heavily loaded with links to slides, videos and photos.
News
The Lean Kanban Central Europe conference in Munich included some excellent presentations, pecha kuchas, and interactive sessions! One of four keynotes, Stephen Bungay, in the final keynote, “Back to the Future” explained how 21st century businesses can learn from 19th century military thinking. Keynotes and select presentations will be available on video soon. Unfortunately not all presentations were recorded, but slides for most are on the LKCE11 website and Interviews with several speakers are up on YouTube. Many thanks to the organizers, presenters and participants who contributed to this fine event - many of whom can be seen in the conference photos at http://www.lean-kanban-conference.de/pictures/). http://www.lean-kanban-conference.de/program/ http://www.youtube.com/user/GroetenUitDelft
We played the” IT Services and Operations” variant of the GetKanban Game simulation at Devopsdays in Gothenborg, Sweden. Participants asked about dependencies between the standard stories and the intangible tickets – hmmm - great idea for the next iteration of the Ops Kanban game! I’d say more, but Gareth Rushgrove (@garethr) already wrote up a lovely summary of the devops conference in his 42nd “Devops Weekly”. Subscribe at http://devopsweekly.com/. http://devopsdays.org/events/2011-goteborg/
Ian Carroll posted a case study on the kanbanops Yahoo! Message board. Included is a nice range of kanban boards from SysAdmin and DBA to Networks and Infrastructure. The metrics (manually captured) come with a warning to management to not use them for performance measurement (love it). http://tech.groups.yahoo.com/group/kanbanops/message/125
The Lean & Kanban 2011 Benelux conference videos are out! The 34 videos might take weeks to watch, but Don Reinertsen and Dave Snowden’s are considered must-watch-now material. Reinertsen’s keynote, “Is it Time to Rethink Deming?” indicates the need to respond to ANY random variation to reduce risk, not just assignable cause variation. Snowden’s keynote, “Practice without Sound Theory Will Not Scale” indicates the need to ABSORB uncertainty rather than reduce uncertainty- otherwise the ability to adapt is destroyed. Enjoy these along with all the other great speakers. http://vimeo.com/channels/leankanban2011benelux
Events
The Lean Enterprise Software and Systems conference (#LESS2011) is happening in Stockholm, Sweden Oct 30 – Nov 2. http://less2011.leanssc.org/
We are packing our bags for Europe this week! With conferences in Sweden and Germany, we anticipate opportunities to meet and talk with some really interesting people - thought leaders of our time!
Yuval Yeret recently blogged on, “How Kanban and TOC Critical Chain Relate”. He discusses the similarities between Multi-Project Critical Chain and Kanban for driving improvement. Yuval points out that while generalization is a worthy vision of Kanban, many organizations simply cannot achieve that (ex: SysAdmins, DBA’s). Therefore, the need to understand the capability of a specialist is key to avoid overburdening them and abusing our scarcest of resources. http://networkedblogs.com/ogomM
Events
Devopsdays (the conference that brings development and operations together) is happening in Goteborg, Sweden Oct 14-15. Am really looking forward to hearing Mattias Jansson and Noa Resare present “A Case Study in Operations and Development Integration at Spotify” and how they used kanban to avoid overloading teams and dealing with “short term panicky stuff”. http://devopsdays.org/events/2011-goteborg/
Next week, I’ll be attending the 1st Lean Kanban Central Europe conference in Munich. Twitter hashtag is #lkce11. Looking forward to keynotes from John Seddon, Kent Beck, David Anderson and Stephen Bungay. And Jurgen Appelo’s is presenting “Complexity Thinking? Or Systems Thinking++?”! But Darn! it’s at the same time as Fridtjof Detzner and Sönke Rümpler’s talk on “Why Kanban Fits the Jimdo Culture”. Good thing the presentations are being recorded:-) http://www.lean-kanban-conference.de/
Please contact .(JavaScript must be enabled to view this email address) with questions.
Lots going on this week with the release of a new book on performance and the Lean & Kanban 2011 conference in Benelux. Discussions aplenty this week around kanban boards reflecting reality (or not) and the impact it has to swimlanes (or not).
Pawel Brodzinski’s recent blog post, “Kanban Story: Kanban Board Should Reflect the
Reality”, hits on a kanban design decision worthy of examination. Pawel discusses the
reasons why a board which reflects something better than reality actually contributes to faulty decisions. http://blog.brodzinski.com/2011/10/kanban-board-reflect-reality.html
A post regarding “Kanban for Network Services team” on the kanbanops board generated considerable discussion from the community this week. Great ideas from contributors on how to handle variability due to different work item task size. And yes, swim lanes were suggested from multiple sources as a means to manage variability. http://tech.groups.yahoo.com/group/kanbanops/message/109
Events
Yet another inspiring AGILEMinds initiative, Lean & Kanban 2011 Benelux hosted by Maarten Volders Oct 3-4 included a line up of amazing speakers and thought leaders from all over the world. http://www.agileminds.be/event/2
See twitter hashtag #lkbe11 for inside scoop from attendees and speakers. Looks like there was some feisty debate over swim lanes and damage capacity. We can’t wait to get a look at all the presentations! For now, David Anderson’s slides on “Kanban - When Is It Not Appropriate” are available at http://agilemanagement.net/images/uploads/KanbanWhenIsItNotAppropriate.pdf
Alan Shalloway’s slides on “Lean Kanban is About People” are available at http://www.netobjectives.com/resources/slides/lean-kanban-about-people-lkbe2011
I joined Daniel Vacanti last week in Seattle for his Kanban training class for Software Engineering. During much of the past year, my focus has been teaching Kanban-for-Ops (Operations) classes, and I had forgotten how the two subjects differ in their curricula.
One of the more significant differences is the inclusion of case studies of ops teams using Kanban.
While teaching ops classes and attending tech conferences in both the U.S. and Europe, I was able to collect a variety of fascinating stories from various technology companies using Kanban in ops.
To date, little has been published on how ops teams implement Kanban, which may explain why the stories are so well received during class. I know a session is going well when the class energy level is sustained right through the mid-afternoon. And nothing, it seems, lifts people’s energy like a real story - eyes light up, voices are raised and questions flow.
On the afternoon of day 2 of last week’s Seattle class, we offered attendees several options:
1) Give more time for people to work on their kanban boards; 2) Dive into the “How to Get Started with Kanban” session; or 3) Look at case studies on Kanban-for-Ops.
The class chose the case studies, which brought a different perspective to how the kanban system that these developers had just designed might be useful downstream. It was fascinating to me to see the interest and excitement this generated. Developers related enthusiastically to the similar problem sets addressed in the ops case studies.
Problems such as conflicting projects, where different teams share limited resources are common. In one case study, requests from “pushy” people attempting to jump their item to the front of the line is handled with a policy requiring justification (via brief business case) during the daily standup meeting. If no objections are raised, the item is allowed to jump the queue. This policy allows for flexibility when needed and reduces annoying attempts to cut in line.
Presenting case studies provides food for thought when designing a Kanban system. They are an excellent way to demonstrate the effect of policies on teams. They also provide attendees with a sense of community – other teams are dealing with similar issues and have found solutions.
Admittedly, all our Kanban training classes and workshops are sprinkled with stories of people using Kanban to solve problems. But in the Kanban-for-Ops training class, case studies are prominent. And if the class surveys are an indicator of value, case studies should remain in the Kanban-for-Ops curricula for some time.
The first time I heard the “DevOps” term was during a conversation with my boss who told me that “DevOps” was the new name for what our Configuration Management team used to do at Corbis. Initially, I took this literally - that DevOps was about source code control systems, building & deploying code. But our team did more than that. We built out and maintained Dev and Test environments. We did Production DB restores. We managed third party dependencies, tools and security access. Searching to learn more about DevOps, I discovered a whole community of IT practitioners engaged in the DevOps movement loosely defined as “bringing Development and Operations together”.
The DevOpsDays conference in Boston was exciting – listening to IT professionals talk about culture as much as they talked about tools! Finally - a movement asking us to think holistically about IT systems development, deployment and operations. I could relate to this movement. Our team did a lot of different things, but we did them by working together with other teams. We worked with Sysadmins to solve environment configuration and domain issues. We worked with DBAs to improve DB backups. We worked with Developers to get their local environment up and running (in a crazy complicated system). We worked with Architects to implement Continuous Integration. We worked with QA to improve lead times. Over time, the nature of our work grew broad in scope and so did our kanban board.
What we really did was build trusting working relationships across the whole IT organization. It was these relationships that helped us resolve many of our previous issues. The Operations teams began attending our monthly Kanban Ops Review where we presented our good, our bad and our sometimes embarrassing results. The transparency brought trust. The vulnerability brought empathy. Sitting in the same room with all eyes on the results helped put us on the same side. The unwanted results became the problem to solve rather than blaming each other.
The DevOps movement appears to be striving towards the patterns in Systems Thinking expressed by W. Edwards Deming. It’s about understanding the system as a whole instead of only looking at the bits and pieces. It’s about work flowing across functions instead of stagnating in silos. It’s about respect and trust amongst individuals doing the work instead of management-driven command and control. The DevOps movement seems to be courageously moving toward an IT service version of Systems Thinking – a truly admirable goal.
Good leadership drives out fear by focusing on the system and can contribute to the success of the DevOps movement. Good leadership combined with a Kanban pull system and trust enables people to do their job with precision and quality resulting in remarkable improvements.
