I’ve got 99 problems, but OpenStack ain’t one – TechReckoning Dispatch v3n8

 OpenStack

It was OpenStack Summit this week, and both suits and t-shirts ate a lot of BBQ in Austin. (Vegetarians, remember there are always the side dishes.)  A few thoughts from an outside observer:

1. If you haven’t at least dipped your toes into OpenStack yet, then it’s not for you right now. Look at who was on stage: Verizon, AT&T. Look at stuff like OPNFV. Does your company have the same problems as AT&T? No? So don’t sweat it.

As OpenStack came up, we had a lot of talk about how OS was not a VMware replacement, and about pets vs cattle. I think implied in that conversation, but often not mentioned explicitly, is a question of scale. Subconsciously, one might have supposed that OS would follow adoption patterns of VMware or AWS: e.g., start with one or two servers, skunkworks projects, and hobbyist tinkering. However, actually operating a cloud is a very different thing. So just get that out of your head: you will not run OpenStack in your normal-sized shop, and OpenStack adoption is not going to be driven by thousands of mid-sized businesses. [Note: There is a future where you do use it, and it might even be installed on premises (think appliances or Platform9), but you ain’t gonna set it up and run it yourself.]

2. I liked Randy Bias’s comment in this video from theCUBE (paraphrased): If you believe in one-size-fits-all for cloud, then public cloud (i.e., AWS) takes over everything. If you believe that one size does not fit all, then there will always be a place for private cloud platforms.

3. OpenStack is estimated to be a $1.2B market growing to $3.4B in 2018, and logically most of that is services. (Caution about revenue estimates in this space: there’s enough creative accounting in “cloud” revenues that IBM alone could report a $1B run rate in crocheted toilet paper cozies if the analysts wanted it.)  I keep thinking about The Software Paradox, which is Stephen O’Grady’s book on how hard it is to sell proprietary software, there’s no wonder why Cisco, IBM, HPE, EMC, Netapp, and Oracle are all there — there’s a metric shit-ton of expensive enterprise software that OpenStack is displacing and (hopefully) expensive hardware that OpenStack will need to work with. The WebSphere or vSphere playbook will not work for a cloud stack in 2016. I still wonder, however, if OpenStack is solving 2010’s problem, not 2020’s. (Note that solving 2010’s problem can be extremely lucrative until 2020 and beyond, because plenty of people are still running Windows Server 2003, thus still having 2000’s problem. Great googly moogly, have you never been in a real IT department?

4. Speaking of which, there’s a certain amount of turtles-all-the-way-down aspect to the current stack. Like vSphere, the OpenStack management engine was really architected to sit on reliable infrastructure. Newer distributed systems like Mesos and Kubernetes were architected to live in a world with more uncertainty. So who runs on the bottom and who runs on the top? For example, do you run OpenStack inside Kubernetes? CoreOS showed off Stackanetes, which does just that. But you can also run Kubernetes insideOpenStack as a container orchestrator. And you can run CloudFoundry, which again has its own distributed resource provisioning and management system, inside the Cisco’s Metapod OpenStack. It makes my brain hurt. I’ve said this before, but who owns the locus of control of this new stack is the defining technology business conflict of our times. If you’re an infrastructure person, it will also affect your career.

Always Read The Comments

Last week we asked about speaking, VMworld, and VMUGs. Sorry if you’re not into VMware — the VMware community really cares about this kind of stuff. I told you VMware has problems other companies only dream about, even in 2016. On more interesting content from vendors, here’s Mat Young:

What I would say as ex fusion-io is that it take talent and money and you have to care. I sadly believe a lot of the tech industry on the vendor side doesn’t care. One question for the audience if vendor A was say 10% more cost but delivered better content and marketing as a service would you buy A over normal B? If you aren’t the decision maker could and would you swing it?

Rob Nelson on bringing new ideas to VMUGs:

I love the idea of more workshops at all conferences. “I solved a problem and would to help you solve the same or similar problem” is probably more appealing to many would-be speakers than being under the bright lights, and probably of interest to attendees. I’ve seen a few do this well, such as CheckPoint at their yearly CPX conference – there’s a lab with some CP software set up and you can run through some suggested labs or just spend time with some product gurus on what interests you the most. More of solving my actual problems, please!

Matt:

1) You can see my thoughts on it here Long story short – I didn’t feel that I had enough to say; I was invited to co-host and jumped at it. I also found a topic that I am familiar with, but also want to learn more about, so I started a local user group. If I can do it, anyone can … they might just need a little push.

2) User talks, or at least user co-hosted talks. The VMware employee sessions are usually good, but the vendor sessions can lapse into a sales pitch. By introducing an end user you all of sudden have real-world problem solving going on. It sort of changes it from a “here’s what we can do” scenario to a “here’s what we did” scenario.

3) Vendors are a necessary evil; that’s not to say that they don’t provide value. I have seen some fantastic vendor sessions, but they seem to be an exception. Sponsorship is needed, unless you turn user groups into a charity or have users pay some sort of membership. Ultimately attendees need to get value out of going. I think that some sort of scoring of vendor presentations would be good. Some ‘sales pitching’ would be acceptable (and necessary), but excessive amounts would result in losing the opportunity to engage your target audience.

Edward Haletky:

Another speaking/presentation alternative if possible is vBrownBag for VMworld, OpenStack Summit, etc. Or just in general off show floors. Be mindful this is a fairly technical audience.

If I was in charge of VMworld talk acceptance I would have a superior, no BS, deep dive technical track. I am not talking 101 or even 201 level elements, which make up the majority of VMworld talks, but truly nitty gritty discussions and answers, a 401 style of talks.  We all want more technology out of VMworld, this is one way to get that.  I would also select have the sessions to deal with security and data protection as just one session is just not going to cut it these days and that is all we really see about security, 1 useful session and a bunch of marketing sessions. I would probably seek more NON-VMware talks than VMware Talks and absolutely none of them be marketing pitches, that is what we get on the show floor. Lastly, I would invite more panels that answer questions from industry experts: The Q&A style discussion goes along way to solving peoples problems.

As for User Groups, I would invite more non-vendor/non-vmware folks to talk even if it is about why they joined the User Group. Bring in mentoring for those who want to talk but need to learn the skills. It can be nerve wracking to get up in front of people.

Luca Dell’Oca:

I stopped a few years ago submitting sessions at VMworld; they never got approved and the answer has always been copy/pasted from a template probably, as it has always been either too generic, in a section where too many others have proposed other sessions, and not linked to VMware tech. I’ve found each of the excuses totally lame, and I checked this by reusing the same presentations at VMUG, where they were all greatly liked. I had the chance to present at both VMworld last year just because my employer is a sponsor and paid for the sessions, so even if again they got high scores, I wouldn’t count them as a success in terms of submission :)

I’ve helped my company [Veeam] to select submissions for our own event, and it was nothing compared to the numbers seen at VMworld. I agree it’s a though job, and sometimes in order to do the first filter, many sessions are declined after a (too) quick look. But somewhere you need to cut when you have 20.000 submisisons and only 500 slots.

I’d like to see more people bringing their sessions to a VMUG if not accepted, but also VMware itself creating one or two tracks for “community” sessions, presented by people who don’t work for a vendor, but they are known in the community. And maybe have the vExpert community vote for them. Not a public vote, as then it would become just a popularity contest, but a vExpert committee would be good.

Justin Paul:

Wow, this was a timely post. I was wrestling with just this question earlier in the week. I think it is important for my career, confidence, and influence (?) to share experiences by getting out there to speak. I submitted a paper for this year’s VMworld (not hopeful), and did last year (not selected), and in 2014 (selected but was co-presenting with a VMware guy). It does seem a bit of a challenge to get strong consideration at the conference, despite the fact that you have some real world, and not always pretty, adventures to share.

Speaking at a large conference is a great experience but I was wondering where else to seek opportunities. I think leveraging user groups is an awesome idea but it seemed that VMUG had been more vendor focused and generally closed off. I’ll have to look more closely at the options there. We try to get our experiences out there however we can (Twitter, blogs, whitepapers) but it would be great to have other conference like options. Something that is also easier and smaller – thinking minor league before the majors. Those smaller events are less intimidating and a good place to get your feet wet. They will also most likely put you in contact with folks that sit more on “your side of the table”. I’m hopeful for VMworld 2016 but the “deck” may be stacked against us – despite having a great story to tell.

From an anonymous correspondent in the ecosystem:

I’ve stopped bothering to submit for VMworld…but not for the reasons you cite. Diane wanted VMworld to be Virtualization World not VMware World, but Diane is long gone.

Are you seriously telling me that talk on containers by someone like Solomon Hikes or Alex Polvi etc. would not put buts in seats? Think about the thousands who attend OpenStack, DockerCon, LinuxCon or Open Network Summit. Don’t tell me this content isn’t relevant to the audience…it is, it’s just threatening to VMware.

VMware of course can chose to fill their conference with any content/speaker they want, it is their conference…but I actually believe that this is short term thinking that actually hurts them in the long run. A miopic focus on Windows nearly killed Microsoft, who is finally learning the lesson by reaching out to the Linux community much more actively.

As long as VMware continues to focus primarily on the products they have on the truck or coming down the pike rather than on the broader set of interests and concerns of their customers the more their customers have to seek that elsewhere. I believe that Diane had it right, by attempting to make VMworld an industry conference she made the market bigger and more diverse and helped build a vibrant ecosystem. Like squeezing one part of a balloon, VMworld’s tight focus on VMware products pushes innovation to other more open venues.

Duncan Epping, Office of the CTO, VMware, thought I was indeed trolling. But I actually was both curious as to people’s thoughts and suspected that some discussion between VMware and the community would be healthy and enlightening.

troll → “fish by trailing a baited line along behind a boat”

Considering you asked a question which you already knew the answer to and you knew would kick off an endless debate whether the VMworld org is doing the right thing or not. As a former insider you know the process inside out, and you know why in some cases things look skewed, you know what the audience rates high and asks for.

But we’ll close with Grant Orchard from VMware giving us insight into the reality of reviewing submissions::

Having been a voter in the session selection process (though not this year) I can tell you that it’s possible for people outside of VMware to get content accepted, but there are a few things that typically prevent them from doing so.
Here are the common mistakes that immediately rule out a session:

1. Title does not match the session abstract, which does not match the session outline.
You need to spend a decent amount of time to create a solid submission. If your submission is sloppy, will your presentation also be that way?

2. Submitting too much in a single session.
You have an hour – if there is no way the content can be presented in an hour then you’re out.

3. Too general.
If 18 people submit a great session on the same topic, only one will get selected. Look for a unique angle on a topic with broad appeal.

4. Being unknown.
This is perhaps more nicely explained as “who are you to speak knowledgeably on this session?”. Engage with VMware product teams through the year. Provide feedback (good and bad) invite them to look at your use cases and build rapport. A lot of voters are from the BU and are rightfully precious about their babies.

Finally, the technical vs marketing aspect. As most people would be aware, it’s hard to gauge how technical a presentation will be. All we have to go on is whether or not a session is marked “technical” or “advanced technical”. Voters are encouraged to vote for more technical submissions, as that is the feedback, but unfortunately you can’t really tell what will be delivered until the day.

Keeping Up

Let me know if you’re in Boston and interested in coming to ZertoCON in Bostn. Use TECHRECKONING15 for 15% off.

One of the Geek Whisperers podcasts from the Silicon Valley VMUG dropped: Investing in Career Insurance: VMware User Group Recap – Episode 110. We’re currently noodling on how to expand peer-based career coaching to other events.

No links this week because we went long on VMUG, but I’ve started sharing daily links again on Twitter, Facebook, LinkedIn, and Google+, which evidently still exists. Check it out.

Just Hit Reply

Is OpenStack on your roadmap? Want to rant more about VMUG? I’m a good listener. Email jtroyer@techreckoniong.com and I’ll include your comments next week.