How are we picking the OpenStack DefCore “must pass” tests?

Fancy ElephantThis post comes with a WARNING LABEL… THE FOLLOWING SELECTION CRITERIA ARE PRELIMINARY TO GET FEEDBACK AND HELP VALIDATE THE PROCESS.
As part of the DefCore work, we have the challenge of taking all the Tempest tests and figuring out which ones are the “must-pass” tests that will define core (our note pages).  We want to have a very transparent and objective process for picking the tests so we need to have well defined criteria and a selection process.
Figuring out the process will be iterative.  The list below represents a working set of selection criteria that are applied to the tests.  The DefCore committee will determine relative weights for the criteria after the tests have been scored because it was clear in discussion that not all of these criteria should have equal weight.
Once a test passes the minimum criteria score and becomes “must-pass” the criteria score does not matter – the criteria are only used for selecting tests. As per the Core principles, passing all “must-pass” test will be required to be considered core.
So what are these 13 preliminary criteria (source)?
1. Test is required stable for >2 releases (because things leaving Core are bad)
  • the least number/amount of must pass tests as possible (due to above)
  • but noting that the number will increase over time
  • least amount of change from current requirements as possible (nova, swift 2 versions)
  • (Acknowledge that deprecation is punted for now, but can be executed by TC)
2. Where the code being tested has an designed area of alternate implementation (extension framework) as per the Core Principles, there should be parity in capability tested across extension implementations
  • Test is not configuration specific (test cannot meet criteria if it requires a specific configuration)
  • Test does not require an non-open extension to pass (only the OpenStack code)
3. Capability being tested is Service Discoverable (can be found in Keystone and via service introspection) – MONTY TO FIX WORDING around REST/DOCS, etc.
  • Nearly core or “compatible” clouds need to be introspected to see what’s missing
  • Not clear at this point if it’s project or capability level enforced.  Perhaps for Elephant it’s project but it should move to capability for later
4A, 4B & 4C. Candidates are widely used capabilities
  • 4A favor capabilities that are supported by multiple public cloud providers and private cloud products
    • Allow the committee to use expert judgement to promote capabilities that need to resolve the “chicken-and-egg”
    • Goals are both diversity and quantity of users
  • 4B. Should be included if supported by common tools (Ecosystem products includes)
  • 4C. Should be included if part of common libraries (Fog, Apache jclouds, etc)
5. Test capabilities that are required by other must-pass tests and/or depended on by many other capabilities
6. Should reflect future technical direction (from the project technical teams and the TC)
  • Deprecated capabilities would be excluded (or phased out)
  • This could potentially become a “stick” if used incorrectly because we could force capabilities
7. Should be well documented, particularly the expected behavior.
  • includes the technical references for others in the project as well as documentation for the users and or developers accessing the feature or functionality
8. A test that is a must-pass test should stay a must-pass test (makes must-pass tests sticky release per release)
9. A test for a Capability with must-pass tests is more likely to be considered must-pass
10 Capabilities is unique and cannot be build out of other must-pass capabiliies
  • Candidates favor capabilities that users cannot implement if given the presence of other capabilities
  • consider the pain to users if a cloud doesn’t have the capability – not so much pain if they can run it themselves
  • “Unique capabilities that cannot be build out of other must-pass capabilities should not be considered as strongly”
11. Tests do not require administrative right to execute
We expect these criteria to change based on implementation experience and community input; however, we felt that further discussion without implementation was getting diminishing returns.  It’s import to remember that all of the criteria are not equal, they will have relative weights to help drive tune the results.

OpenStack Core Definition (DefCore) Progress in 6 key areas

DefCore Elephant Cycle

I’m excited to report about the OpenStack Board progress on defining OpenStack core.  At the Hong Kong summit, Joshua McKenty and I were asked to chair a new standing committee, now known as DefCore, to define “OpenStack Core” based on the core principles that we determined over the last 6 months (aka “the spider”).

Joshua and I took on the challenge with gusto and I’m proud to say that we’ve already made significant progress against an aggressive timeline to have the pilot must-pass tests for Havana defined before the Juno Summit in April 2014.  It’s important to remember that we’re moving from a project based definition of core to test-driven capabilities because this best addresses our interoperability objectives.

In the 8 weeks since the summit, we’ve had six very productive meetings (etherpads for Prep, DefCore.1, DefCore.2, Criteria 1 and 2) with detailed notes and recorded content. Here’s my summary of our results so far:

  1. An Aggressive Timeline for having pilot Havana must-pass tests approved by the Juno summit in May 2014.  That drives the schedule backward toward a preliminary list in March.  Once we have a pilot list for Havana, we expect to have Ice House done +90 days and Juno done at the Paris summit.

  2. Test Selection Criteria a preliminary set of 14 criteria (needs a stand alone post) that will be used to quantitatively score the current 700+ tests.  We also agreed to use a max 100 point weighting system for the criteria.  The weights and score requirement iteratively once we have done a first scoring pass.  Our objective is to make must-pass test selection as objective and transparent as possible (post with details).

  3. Distinction between Capability & Test is important because we recognize that individual tests may validate multiple capabilities and individual capabilities may have multiple tests.  Our hope is to present the results in terms of capabilities not individual tests.

  4. Holding Off on Bylaws Changes needed to clarify how OpenStack manage core definition.  It was widely expected that the DefCore committee would have to make changes to the OpenStack bylaws; however, we believe we can proceed without rushing changes.  We have an active subcommittee preparing changes in advance of the next DefCore cycle.

  5. Program vs. Project Definition efforts are needed to help take pressure off requests to have “projects promoted to core status” and how the OpenStack trademark is used for projects.  We are trying to clarify OpenStack Programs (e.g.: OpenStack™ Compute) carry to the trademark while OpenStack Projects (e.g.: Nova and Glace) are members of those programs and do not carry the OpenStack trademark directly.  Consequently, we’d expect people to say “OpenStack Compute Project Nova” instead of “OpenStack Nova.”  This approach addresses several issues that impact DefCore Board activities around trademark, core and brand.

  6. RefStack Development and Use Cases provide the framework for community reporting of test results.  We consider this infrastructure critical to getting community input about must-pass tests and also sharing interoperability information.  This effort is just beginning and needs help from the community.

For all this progress, we are only starting!  We’ve cleared the blocks preventing implementation and that will expose a new set issues to discuss.  Look for us to start applying the criteria to tests in the next months.  That will quickly expose the strengths and weaknesses of our criteria set.  We’ve also got to make progress on Program vs. Project and start RefStack coding.

We want community participation!  Please let us know what you think.

Mark Stouse’s “Making Predictions for 14″ series

I was invited to be part of Mark Stouse’s 2014 big data & cloud predictions series.  His questions had me thinking deeply about the past year and I’m happy to repost them here with links to the other predictors too including (Robert ScobleShel Israel, and David H. Deans).

1.  Describe in one sentence what you do and why you’re good at it.

I specialize in architecture for infrastructure software for scale data center operations (aka “cloud”) and I have 14 years of battle scars that inform my designs.

 2.  Cloud Computing, Big Data or Consumerization: Which trend do you feel is having the most impact on IT today and why?

Cloud, Data & Consumerization are all connected, so there’s no one clear “most impactful” winner except that all three are forcing IT to rethink how we handle operations.   The pace of change for these categories (many of which are open source driven) is so fast that traditional IT governance cannot keep up.  I’m specifically talking about the DevOps and Lean Software Delivery paradigms.  These approaches do not mean that we’re trading speed for quality; in fact, I’ve seen that we’re adopting techniques that deliver both higher quality and speed.

 3.  What do you think is the biggest misconception about Cloud computing/Big Data/Consumerization?

That someone can purchase them as a SKU.  These are really architectural concepts that impact how we solve problems instead of specific products.  My experience is that customers overlook their need to understand how to change their business to take advantage of these technologies.  It’s the same classic challenge for ROI from most new technologies – they don’t exist apart from the business matching changes to the business to leverage them.

 4.  Which (Cloud Computing/Big Data/Consumerization) trend has surprised you most in the last five years?

Open source has surprised me because we’ve seen it transform from a cost concern into a supply chain concern.  When I started doing open source work for Dell, customers were very interested in innovation and controlling license costs.  This has really changed over the last few years.  Today, customers are more concerned with community participation and transparency of their product code base.  This surprised me until I realized that they are really seeking to ensure that they had maximum control and visibility into their “IT Supply Chain.”   It may seem like a paradox, but open source software is uniquely positioned to help companies maintain more control of their critical IT because they are not tightly coupled to a single vendor.

 5.  How has Cloud Computing/Big Data/Consumerization had the biggest impact in YOUR life to date?

Beyond it being my career, I believe these technologies have created a new degree of freedom for me.  I’m answering these questions from the SFO airport where I’m carrying all of the tools I need to do my job in a space small enough to fit under the seat in front of me plus a free Wifi connection.  I believe we are only just learning how access to information and portable computing will change our experience.  This learning process will be both liberating and painful as we work out the right balances between access, identity and privacy.

 6.  On a lighter note – If Cloud/Big Data/Consumerization could be personified by a superhero, which superhero would it be and why?

The Hulk.   Looks like a friendly geek but it’s going to crush you if you’re not careful.

 7.  What aspect of (Cloud Computing/Big Data/ Consumerization) are you most excited about in the future, and what excites you about it?

The Internet of Things (even if I hate the term) is very exciting because we’re moving into a place where we have real ways to connect our virtual and physical lives.  That translates into cool technologies like self-driving cars and smart power utilities.  I think it will also motivate a revolution in how people interact with computers and each other.  It’s going to open up a whole new dimension on our personal interaction with our surroundings.   I’m specifically thinking about a book “Rainbows End” by Vernor Vinge that paints this future in vivid detail.

Competition should be core to OpenStack Technical Meritocracy

In my work at Dell, Technical Meritocracy means that we recognize and promote demonstrated talent into leadership roles. As a leader, one has to make technical judgments (OK, informed opinions) that focus limited resources in the (hopefully) right places. Being promoted does not automatically make someone right all the time.

I believe that good leaders recognize the value of a diverse set of opinions and the learning value of lean deliverables.

OpenStack is an amazingly diverse and evolving community. Leading in OpenStack requires a level of humility that forces me to reconsider my organization hierarchical thinking around “technical meritocracy.” Instead of a hierarchy where leadership chooses right and wrong, rising in the community meritocracy is about encouraging technical learning and user participation.

OpenStack is a melting pot of many interests and companies. Some of them naturally aligned (customers+vendors) and others are otherwise competitive (vendors). The vast majority of contribution to OpenStack is sponsored – companies pay people to participate and fund the foundation that organizes events. That does not diminish our enthusiasm for the community or open values, but it adds an additional dimension

If we are really seeking a Technical Meritocracy, we must create a place where ideas, teams, projects and companies can pursue different approaches within OpenStack. This is essential to our long term success because it provides a clear way for people to experiment within the project. Pushing away alternate approaches is likely to lead to forking. Specifically, I believe that the mostly likely competitor to any current OpenStack project will be that project’s .next version!

Calls for a “benevolent dictator” imply that our meritocracy has a single person with perspective on right and wrong. Not only is OpenStack simply too complex, I see our central design tenant as enabling multiple approaches to work it out in the community. This is especially important because many aspects of OpenStack are not one-size-fits all. The target diversity of our community requires that we enable multiple approaches so we can expand our user base.

The risk of anointing a single person, approach or project as “the OpenStack way” may appear to streamline the project, but it really stifles innovation. We have a healthy ecosystem of vendors who gladly express opinions about the right way to implement OpenStack. They help us test OpenStack technical merit by finding out which opinions appeal to users. It is essential to our success to enable a vibrant diversity because I don’t think there’s a single right answer or approach.

In every case, those vendor opinions are based on focused markets and customer needs; consequently, our job in the community is to respect and incorporate these divergent needs and find consensus.

Spinning up OpenStack “DefCore” Committee by spotting elephants

ElephantsThis week, Joshua McKenty, me and a handful of interested individuals (board member Eileen Evans included) met to start organizing the DefCore* Committee.  This standing committee was established by an OpenStack Foundation resolution just before the Hong Kong Summit.  Joshua and I were nominated as co-chairs (and about half the board volunteered to be members).    This action was an immediate result of the unanimous passage of the 10 Principles that I was driving in the DefCore “Spider” cycle.

We heard overwhelmingly at the Hong Kong summit that defining core should be a major focus for the Board.

The good news is that we’re doing exactly that in CoreDef.  Our challenge is to go quickly but not get ahead of community consensus.  So far, that means eating the proverbial elephant in small bites and intentionally deferring topics where we cannot find consensus.

This meeting was primarily about Joshua and I figuring out how to drive DefCore quickly (go fast!) without exceeding the communities ability to review and discuss (build consensus!).  While we had future-post-worthy conceptual discussions, we had a substantial agenda of get-it-done in front of us too.

Here’s a summary of key outcomes from the meeting:

1)      We’ve established a tentative schedule for our first two meetings (12/3 and 12/17).

  1. We’ve started building agendas for these two meetings.
  2. We’ve also established rules for governance that include members to do homework!

2)      We’ve agreed it’s important to present a bylaws change to the committee for consideration by the board.

  1. This change is to address confusion around how core is defined and possibly move towards the bylaws defining a core process not a list of core projects.
  2. This is on an accelerated track because we’d like to include it with the Community Board Member elections.

3)      We’ve broken DefCore into clear “cycles” so we can be clearer about concrete objectives and what items are out of scope for a cycle.  We’re using names to designate cycles for clarity.

  1. The first cycle, “Spider,” was about finding the connections between core issues and defining a process to resolve the tension in those connections.
  2. This cycle, “Elephant,” is about breaking the Core definition into
  3. The next cycle(s) will be named when we get there.  For now, they are all “Future”
  4. We agreed there is a lot of benefit from being clear to community about items that we “kick down the road” for future cycles.  And, yes, we will proactively cut off discussion of these items out of respect for time.

4)      We reviewed the timeline proposed at the end of Spider and added it to the agenda.

  1. The timeline assumes a staged introduction starting with Havana and accelerating for each release.
  2. We are working the timeline backwards to ensure time for Board, TC and community input.

5)      We agreed that consensus is going to be a focus for keeping things moving

  1. This will likely drive to a smaller core definition
  2. We will actively defer issues that cannot reach consensus in the Elephant cycle.

6)      We identified some concepts that may help guide the process in this cycle

  1. We likely need to create categories beyond “core” to help bucket tests
  2. Committee discussion is needed but debate will be time limited

7)      We identified the need to start on test criteria immediately

  1. Board member John Zannos (in absentia) offered to help lead this effort
  2. In defining test criteria, we are likely to have lively discussions about “OpenStack’s values”

8)      We identified some out of scope topics that are important but too big to solve.

  1. We are calling these “elephants” (or the elephant in the room).
  2. The list of elephants needs to be agreed by DefCore and clearly communicated
  3. We expect that the Elephant cycle will make discussing these topics more fruitful

9)      We talked about RefStack code features

  1. Allowing users to upload/post test results from their clouds to enable white box test reporting
  2. Allowing users who have uploaded results to +/- vote on tests they think are important
  3. We established a requirement that posting results requires an OpenStack ID
  4. We established a requirement that only a single Corporate designate (provided by the Foundation) can make a result official for their company.
  5. Collecting opt-in data with test results using tags for things like alternate implementations use, host operating system(s), deployment method, size of cloud, and hypervisor.
  6. We discussed (but did not resolve) that it could be possible to have people run RefStack against public cloud end points and post their results
  7. We agreed that RefStack needs to be able to run locally or as a hosted site.

10)   We identified a lot of missing communication channels

  1. We created a DefCore wiki page to be a home for information.
  2. Joshua and I (and others?) will work with the Foundation staff to create “what is core” video to help the community understand the Principles and objectives for the Elephant cycle.
  3. We are in the process of setting up mail lists, IRC, blog tags, etc.

Yikes!  That’s a lot of progress priming the pump for our first DefCore meeting!

* We picked “DefCore” for the core definition committee name.  One overriding reason for the name is that it has very clean search results.  Since the word “core” is so widely used, we wanted to make sure that commentary on this topic is easy to track against the noisy term core.  We also liked 1) the reference to DefCon and 2) that the Urban Dictionary defines it as going deaf from standing too close to the speakers.

5 differences between Cloud ops and Bare Metal ops

OpenStack SummitCloud APIs are about abstracting operations to simplify deployment.  We want users of our cloud infrastructure to operate with blissful unawareness of the underlying networking topology, storage configuration and physical infrastructure.  For their perspective, the cloud is perfectly elastic, totally configurable and wonderfully consistent. Cloud Admins on the other hand need visibility and controls that expose the complexity while keeping it rational. These are profoundly different concerns.

Maintaining the illusion of clean and simple Cloud ops infrastructure is very valuable; however, it’s just an illusion.  The black metal box behind those APIs is complex, messy, unpredictable and dynamic.

1. Metal Ops has to deal with network topology and details like if an operating system enumerates the NICs correctly, bonding the correct NIC pair and which 10g network to use for the storage traffic. In networking, the topology determines how much traffic you can subscribe to a link and how to provide resliency. Networking does not exist in isolation: you must consider the boundary firewalls and routers to either block or allow traffic because without connectivity the cloud is useless. Details like the access and registration in DNS, NTP and DHCP provide foundations our stable operations. These details are (and should be) hidden from the cloud user.

2. Metal Ops has to deal with firmware issues at every level.  It matters to the server if it boots into BIOS or UEFI mode.  We have to manage the fact that RAID partitions need to be optimized based on the workload and type of drive.  We have to consider if there are specialized drivers and caches to manage and security features (like Intel TXT) to activate.  These details are (and should be) hidden from the cloud user.

3. Metal Ops have to consider the security of their infrastructure.  We have to manage where the admin control network crosses security domains.  It matters which layer 2 networks have access to which parts of the infrastructure.  Separation of responsibility for network vs. storage vs. compute is a reality that it not going away. These details are (and should be) hidden from the cloud user.

4. Metal Ops have to manage operating system compatibility.  I know personally that vendors test and certify their operating systems on an enormous matrix of silicon.  I also have learned that the matrix of possible combinations is far larger and fundamentally impossible at the edges.  There’s a reason that operators seek homogeneous environments and LTS releases. These details are (and should be) hidden from the cloud user.

5. Metal Ops have to deal with hardware failures. By simple statistics, the larger the system the more things will break and metal ops have to cope with this reality. We have to expose failure zones and boundaries to make intelligent responses (like moving data from a failed drive to a non-adjacent one) that require intimate knowledge of system topography that are intentionally hidden in cloud ops. Further, we have to have monitoring and management tooling that knows how to identify which NIC in a bond failed or flash the lights on the failed drive of an array. These details are (and should be) hidden from the cloud user.

Cloud’s power is being able to abstract away this complexity.  Dealing with it gracefully behind the scenes requires transparency and details that make Metal Ops job fundamentally different.

While both can be highly automated and pass my “Cloud is Infrastructure with an API” test, their objectives are different.

Crowbar HK Hack Report

Purple Fuzzy H for Hackathon (and Havana)Overall, I’m happy with our three days of hacking on Crowbar 2.  We’ve reached the critical “deploys workload” milestone and I’m excited about well the design is working and how clearly we’ve been able to articulate our approach in code & UI.

Of course, it’s worth noting again that Crowbar 1 has also had significant progress on OpenStack Havana workloads running on Ubuntu, Centos/RHEL, and SUSE/SLES

Here are the focus items from the hack:

  • Documentation – cleaned up documentation specifically by updating the README in all the projects to point to the real documentation in an effort to help people find useful information faster.  Reminder: if unsure, put documentation in barclamp-crowbar/doc!
  • Docker Integration for Crowbar 2 progress.  You can now install Docker from internal packages on an admin node.  We have a strategy for allowing containers be workload nodes.
  • Ceph installed as workload is working.  This workload revealed the need for UI improvements and additional flags for roles (hello “cluster”)
  • Progress on OpenSUSE and Fedora as Crowbar 2 install targets.  This gets us closer to true multi-O/S support.
  • OpenSUSE 13.1 setup as a dev environment including tests.  This is a target working environment.
  • Being 12 hours offset from the US really impacted remote participation.

One thing that became obvious during the hack is that we’ve reached a point in Crowbar 2 development where it makes sense to move the work into distinct repositories.  There are build, organization and packaging changes that would simplify Crowbar 2 and make it easier to start using; however, we’ve been trying to maintain backwards compatibility with Crowbar 1.  This is becoming impossible; consequently, it appears time to split them.  Here are some items for consideration:

  1. Crowbar 2 could collect barclamps into larger “workload” repos so there would be far fewer repos (although possibly still barclamps within a workload).  For example, there would be a “core” set that includes all the current CB2 barclamps.  OpenStack, Ceph and Hadoop would be their own sets.
  2. Crowbar 2 would have a clearly named “build” or “tools” repo instead of having it called “crowbar”
  3. Crowbar 2 framework would be either part of “core” or called “framework”
  4. We would put these in a new organization (“Crowbar2″ or “Crowbar-2″) so that the clutter of Crowbar’s current organization is avoided.

While we clearly need to break apart the repo, this suggestion needs community more discussion!