Blogs

RedMonk

Skip to content

IBM Pulse 2011 – The Tivoli with two minds – Trip Report

Steve Mills talking up Smarter Planet at IBM Pulse

  • Pulse is two conferences: bow IBM helps companies manage computers and how IBM helps companies manage the world’s infrastructure. Bridging those two can be difficult.
  • Often, IBM emphasizes the end result and the “big picture” solution instead of showing demos and speaking to exact technologies used to solve the big problems. Balancing both establishes better credibility than just doing one or the other.
  • IBM gave insight on how cloud sales are going: 1,000’s of engagements, mostly private cloud, and, IBM claims, non-x86 systems (which IBM has to sell) can be a better fit.
  • For the first time, IBM started talking about “dev/ops,” speaking to the benefits of having development and operations work more closely together moderated by new technologies and tools. They’re looking towards OSLC and some in-beta provisioning and image management software to help.
  • Tivoli has been doing much work beyond the data center, helping utilities and other companies manage their non-computing infrastructure.

In recent years, Tivoli has been of two minds: classic Tivoli managing data centers and IT, and then the new Tivoli looking to manage the world’s infrastructure, not just its computers. After picking up MRO Solutions and going all “Smart Planet” at the corporate-level, Tivoli speaks much more in public about waste water, building, and city management. One presumes that there’s a better market in managing the “Smart Planet” than there is in just the data center.

The annual Pulse conference, then, always leaves some part of the audience flummoxed asking who this conference is for: sysadmins, building managers, IT bosses, power companies, BSM dashboarders, or utilities? As you can imagine, it drives us IT analysts crazy: we’re used to working on narrow products and lines of businesses – at the very least, technologies. IBM instead wants us to take a more “big picture,” solution-driven approach to looking at them: ithe details of the exact customer use of IBM (and partners!) matters less than the fact that IBM was involved and played a major role in improving the business.

Digging into the Smart Planet

Just as a math teacher would demand of you, a company must show its work when it comes to impressive solutions.

Schiphol is not really buying software. They’re buying performance and we took up the challenge of IBM to design and deliver performance.
Amsterdam Airport Schiphol Case Study Video

IBM has a bit of a presentation problem, really with this: they want so much to talk “high level” and not get into the gorpy details of the technologies used. They have the exact opposite of the problem Microsoft has: all technology and product with not enough business. The thing is, as a technology company you have to open up the can of worms and spill it out to gain credibility: we trust that you can assemble technologies together to solve big problems only if you tell us about the technologies. The more high-level and the less technology based, the vaguer it is and the harder it is to take seriously without a lot of extra leg-work hunting down those missing items.

For any given Smarter <insert noun with a big budget here>, the technology story is there and should be easy to tell. In Tivoli land it’s this: imagine if everything had an IP address and was sending and receiving data on a network of a bunch of “things” that are all working together towards some goal(s), just like a computer network. Garbage trucks, rooms in buildings, pipes in utilities, and power meters on homes – all of these things (can) exist in networks now, and Tivoli knows how to manage networks of things (starting with computers). Here’s some screenshots and demos of that management in action…

Structurally, that’s just like a data center problem: lots of data flying around that you need to track; domain knowledge to understand what the data is and means in itself and in the overall context; analytics to report over the state of things that let you decide what to do next; and, eventually, a fancy dashboard with the Three Colors (Green, Yellow, Red; Good, Get Ready to freak out, FREAK OUT!). This pattern applies to computer networks, to telco networks, clouds, buildings, city-as-networked-system, and whole planet-systems if you’re lucky enough to get that contract.

Instead of having to be Indian Jones to ferret this out, it’d be great to just see Mr. DNA explaining it, gnarly SKUs and all. If dig even shallowly into the details of the oft cited DC Water win, there’s plenty of fascinating technology details. Imagine what you could show with a 10 minute demo in a keynote!

Products, Statements, and Meaty Propositions

More analyst action at #ibmPulse

Nonetheless, amongst the passionate stories about running Swiss trains on time and tracking beef from field to feast, there were some definitive pointers to what you’d expect, technologies:

  • BSM – A Business Service Management demo showed how “everything” is pulled together to manage a theoretic airport (I don’t think they realized that “airport” and “BSM” were all but trade-marked by BMC back in the BSM heyday) – it was very nice and much welcome as, really, the only demo in the keynotes.
  • Product Announcements – A rapid-succession of bullet points on the second day keynote slide introduced a slew of new and updates IT Management offerings. This press release seems to round them up, including a new beta offering for cloud management that looks nice (but that’s difficult to find outside of the press release).
  • 1,000’s of cloud customers – In that same press release, there’s some “cloud wins” chest thumping: “IBM has helped thousands (!!) of clients adopt cloud models and manages millions of cloud based transactions every day in areas as diverse as banking, communications, healthcare and government, and securely tap into IBM cloud-based business and infrastructure services.”
  • Private Cloud is where the money is – And when it comes to those clouds, IBM said time and time again that the revenue was all in private cloud at the moment. One snarky analyst said that when faced with the notion from IBM that their mainframe (System z) is the best way to run a “cloud” he likes to challenge them to start charging Amazon prices. Indeed, at several points, claims were made that a z was at least comparable in price to cloud pricing – be a nice study to hunt down. (In the meantime, check out this RedMonk overview/interview of the newest z IBM has available.)
  • How do companies decide on public vs. private cloud? “What we came to see: if that processes that are not critical to my business, [companies] don’t want that in the public cloud,” IBM’s Robert LeBlanc said, instead they want private cloud.
  • What is a “private cloud”? But then what makes for a private cloud? As Tom Rosamilia said in another part of the analyst sessions: the cloud ends up giving us a new method of delivery and what people charge for [those IT services].
  • Image Management – When you look at how IBM would like to do cloud, they’re largely (only?) still on images instead of the model-driven automation (see one discussion via Amazon CloudFormation here) we see from Puppet and Chef.
  • Warming up to dev/ops – There was a strong (enough) theme of dev/ops: the word was uttered much. While there weren’t really products ready to go to support a dev/ops like way of delivering software, there was much speaking to it. “It’s time for ops to insert itself and really be heard,” Harish Grama said.
  • The Developer Land-grab – During an executive panel at the analyst event, there was much “dev/ops love” as it were. Neeraj Chandra even spoke to the developer land-grab saying that “not so long ago, QA was the enemy” and in the same way that they were brought into the team, it seems like getting operations involved as part of the overall team is a good idea.
  • My advice to IBM was to get out in front of this dev/ops now – they had a tragically long lag time with cloud, and there’s a tremendous amount of room in the enterprise-y space to get at least one IBM “evangelist” out there talking about dev/ops and cloud at a technical, developer friendly level.
  • A Standard for dev/ops – IBM really wants OSLC to work out as a model to do all of this. Indeed, the work of the likes of TaskTop of Rational’s Team Concert can be impressive. When asked by one analyst about getting more industry heavy-weights involvement to make it standardized, it was pointed out that Oracle is involved, Microsoft won’t ever join anything (the implicit rhetorical question being: so who cares?), and that HP should sign-up. See the sort of “how we’re doin'” dashboard for the group for more.
  • Get rid of your IT – The idea of consolidation was a sort of theme through-out: consolidation onto the higher end boxes IBM has to sell (POWER, System z [mainframe], and x if you really want that). In a characteristically bombastic and plain-spoken keynote, Steve Mills summed it up, as in here: “The fewer boxes you have means the fewer power supplies, which means less energy and less labor. So doing more computing with fewer boxes is a good thing fundamentally. You can’t deny it. It’s arithmetic. Anybody who thinks otherwise is not thinking clearly.” Later on, he added: “Now, obviously, if you’re not a company that makes one of these, you don’t like them. So, if you’re only in the Intel server business, you don’t like these things and you have a lot of disparaging things to say about these systems. Duh, of course that’s what HP is saying, of course that’s what Dell is saying, of course that’s what Sun/Oracle is saying, of course they’re saying that. Duh, doesn’t mean you have to listen to it. People say stupid crap every day.” You can put that last sentence on a button and start wearing it at conferences. (Also, see TPM’s recent overview of IDC’s server market-share estimates.)

More

Categories: Companies, Conferences, Enterprise Software, Systems Management, The Analyst Life.

Tags: , , , ,

Comment Feed

9 Responses

  1. Poor IBM. And HP and CA and BMC really. They're struggling to address these disruptive areas with very un-disruptive ideas. I tried to read the OSLC stuff but my colon started to spasm, it's almost as opaque as Microsoft's Azure marketing. Maybe an old Tivoli hand like John Willis can translate this into "normal devops-ese" for us.

  2. Hi Ernest, I participate on OSLC on behalf of IBM Tivoli and would be very interested in any (constructive) feedback you could provide on how we could make OSLC easier to understand. I agree with you it's not nearly as approachable as it could be. We plan to work to improve this over the course of the year.

    FYI, VMWare recently announced that they will be using OSLC as part of their Code2Cloud initiative: http://www.vmware.com/company/news/releases/sprin

    This was a nice surprise since at least I didn't know they were looking at OSLC until people starting tweeting about that press release.

  3. Hey Bill. Well, the Youtube video helps, it explains the goals of the project pretty well. But the verbiage on the site itself is very opaque. I do have some constructive ideas. Try restating it in English. Use bulleted lists and examples. Have "paths in" for different user types (tool builder? tool consumer? OSLC developer?) that channelizes people to relevant content. There's "why collaborate" real high level content next to inexplicable detailed specs in one big list.

    Example restatement:
    "a community and set of specifications for Linked Lifecycle Data" – I don't have a clear idea what that means and as a result neither does 99.9% of the human race. Even using ALM without explanation is only useful to those who have too much spare time in their lives and as a result read Gartner/Forrester/etc reports.

    "a set of RESTful standards so that all your requirements, change tracking, bug reporting, asset tracking, build, and other tools can talk to each other" is better though kinda begs the question of "what, like CMDBs promised but didn't deliver on?"

    Like many others, I have somewhat abandoned the large enterprise frameworks and heavy processes, and three letter acronyms in favor of more of a Visible Ops and working DevOps tools (e.g. chef, puppet) approach. I do think there's a need for a "toolbelt for the toolchain" but whatever that is has to be directly relevant to tool builders and users and the communication needs to hook people in. Depending on who you're looking to target, that is.

  4. Hi Ernest. Thanks for taking the time to respond. I agree with most of your sentiments – and many of the changes to the OSLC web site you suggest align with our plans. The OSLC specifications themselves take a minimalist approach to integration, so it's unfortunate that the site's text gives you a different impression. We (the OSLC community) will fix it.

    Also, we too are not fans of three letter acronyms. That's why we chose the name "Open Services for Lifecycle Collaboration" – it produces a four letter acronym.

    ;-)

  5. Cool. I am in general interested in the concept – one of the things we’re doing in house is creating a system model that we conceive of as a “toolbelt” for the devops “toolchain” because, as the YouTube video notes, there’s all these tools that want you to model in them/interact with them separately and that’s a big pain point – but 18 years in IT has made me a twitch monkey; if some standard doesn’t make sense in the first 5 minutes then it’s unlikely to get wide adoption, and without wide adoption any standard is just limitations on your software with no recompensating value. I like to tell people “there’s the right way, the wrong way, and the standard way” so am all about picking up good and relevant standards… But the sad thing about standards sometimes is that there are so many.

  6. I really like the “must make sense in five minutes” rule of thumb – brilliant – and will set as a goal.

    Thanks again very much for your input.

    PS – I found your blog and really enjoyed the “Scrum for Operations” entry (http://bit.ly/fX6RKd for other folks). Until recently I worked on a dev team (Rational Jazz development) where we followed a pretty rigorous Scrum approach and built tooling for Scrum, so it was really interesting to think about applying Scrum principles to operational tasks.

  7. Michael,
    I’m sorry I missed you at Pulse this year. I would’ve liked to have seen more pics of your adorable little one. :)

    Thanks for taking the time to provide your write up. You captured many of the same things I was feeling, which is why I conducted a wrap up interview with Scott Hebner (http://ow.ly/1bAqf1).

    As I think about other conferences I attend, I expect the keynotes to be pretty strategic–often in a way that goes way beyond what I was thinking of in my little old day job. Do you think this is the case with such conferences as Pulse too? If not, what would be an ideal agenda in the main tent at such a conference?

Continuing the Discussion

  1. […] Joyent, or some other based cloud?), at the moment it’s better than the alternative of not being technical enough or just choosing one stack. Still, in the near future, Dell will need to stream-line, or at least […]

  2. […] Kuttler to discuss the happenings at IBM’s Tivoli conference, Pulse. Also, be sure to check my trip report from this year’s Pulse – there’s some good discussion in the […]