Blogs

RedMonk

Skip to content

IT Management Podcast #004 – Cloud-crazy!

Download the episode directly, or subscribe to the podcast feed.

This week we start out with a review of John’s monitoring panel at barcampESM. As he’s said in other forums, he was pleased with the result. I then mention announcements and whitepapers around the Common Model Library (CML), which is a further evolution of the SML family of IT Management data models. There’s a large cross-vendor effort, similar to the CMDBf, but there doesn’t seem to be any open source folks – who knows if they weren’t invited, or weren’t interested.

We move into out cloud talk episode of the episode, with me mentioning that I’m seeing the emergence of a bunch of “EC2 Juniors” sprouting up, like Rackspace’s recent virtualization announcement. John tells us he’s been digging into cloud talk quite a lot – into Mosso, Rackspce, and 3Tera.

We touch on 3Tera briefly, who John talked with recently. Recalling a past briefing with them, I explain that they’re basically cloud-in-a-box software that allows you to build your own grid, or build it out somewhere else. John then clarified that Rackspace’s offering is just to run your VMWare server, not quite the same as running a cloud for you. Mosso, on the other hand, has a more virtualized, grid-driven setup.

I ask John what he seems people running in the cloud – what type of applications. So far it seems like public web site applications like WordPress, drupal, and web servers. Before digging too much more into that discussion – which we pick up later – we dig into 3Tera more. 3Tera creates and sells the software to run a grid along with the management console for setting up and tying together components in the grid. You buy the software, and either install it in your own data center, or one of the data center providers that 3Tera works with. There’s lots of drag-n-dropping to combine together load-balancers, databases, and web services.

After John’s detailed discussion of what 3Tera does, I jump back to the discussion of what people will run on these grids – what “work-loads” people can move to it. I re-cap the briefing Stephen O’Grady and I had a while ago with 3Tera and the frustrating we had around this question. We were thinking, sure, this grid stuff sounds great, fantastic. But, let’s say we run an SAP install on-top of it, something goes wrong, we call up SAP support, and the first thing they ask us is “what operating system are you running it on?” If we tell them it’s some grid technology they’ve never heard of, we’ll probably get the support boot.

The point is, when it comes to enterprise, business software, there’s a lot of work to be done now to get existing business software to run, supported on all this new cloud stuff. Currently it seems to me we’ve got great technologies for running web site stacks and infrastructure for ISVs building out their own software. But for business users, for “enterprises” running other people’s software, there’s a huge gap in the glue-tooling between existing business software and being able to run it “in the cloud.” We don’t have any idea what this would look like, whether it’s one of “the children of the VNC” type applications of what, but there doesn’t seem to be anyone working on the problem.

My suggestion, of course, is that this is a chance for a business or two: a framework that retrofits existing software to run in the cloud. Sure, the “real” solution is for software companies to write their new software “grid native,” but that’ll take a long time. Check out EnterpriseDB’s cloud edition as well.

John points out that this retooling could accelerate if Wall Street finally gets wise to the cost savings available by running stuff in the cloud. His premise is that there’s a waste in the duplication of running data-centers, on-premise things. But, if investors got wind of how much savings were available – if Mad Money Jim Cramer were yelling about it and pressing red bonkers-sound buttons – the IT world would figure it out right quick. We joke that this would be “the ultimate business/IT alignment.”

The reward, as we get into, is the promise of cheaper and easier to run IT. On the face of it, this means less people. While good for “business,” bad for those people who get laid of. I ask John, “what about the IT guys out of jobs?” and as he points out, technology has always seemingly reduced jobs and at the same time required lots of people to run. That is, it’ll probably be all right. More specifically, by way of anecdote, John says there’s so much “busy work” in IT now-a-days, that sopping up that busy work – like getting a developer Oracle instance spun up – is the real goal, which would free up people to do more important work, which there’s no lack of.

I then ask John what he meant by an earlier comment along he lines of ESM not going anywhere. He clarifies that he means nothing much is going to change in ESM, and then tells us about Doug McClure’s idea for a Systems Management Database. Essentially, a unified console and central “brain” that sucks in monitoring data from all sorts of different agents, devices, and everything else – a layer above everything else that creates on place to look. While this sounds like what ESM is supposed to do in the first place, the slight difference that I glen is that the SMDB is supposed to unify the fragmented groups and tools that exist in IT shops. Rather than assume one tool will do away with those different silos, it instead accepts them and provides a new view of them.

Out of the cloud and friends, John asks me about the possibility of Microsoft/Yahoo! now that Microsoft has an extended an offer to buy. I tell him the results of my Twitter poll, asking if people thought Microsoft would do right by flickr and del.icio.us. Pretty much everyone replied that they were worried that Microsoft would mess it up. I point out that it’d introduce a whole lot of new technology and cultures to Microsoft that Redmond wouldn’t have brought on itself otherwise: OpenID, LAMP-like stacks for hardware, and general non-Microsoft IT. (Also, see another RedMonk take from James Governor.)

We round up the the episode by talking about the recent Hyperic release, touching on performance fixes and Nagios importing. I note that it seems like all of the open source IT management platform folks are gearing up their performance chops to go for the enterprise management space rather than just the mid-market they’re ostensibly known for. On Nagios, I paint out that the Nagios importing could enable either replacing or working with Nagios instals.

Finally, John asks about the RedMonk 5th Birthday party next week in SFcome on by for a drink if you like! And then he points out RedMonk’s recent award as part of LinuxWorld’s 2008 Open Source Business Leaders series.

Disclaimer: see the RedMonk clients list for a RedMonk clients mentioned in the podcast.

Categories: IT Management Podcast, Podcasts, Systems Management.

Comment Feed

2 Responses

Continuing the Discussion

  1. […] IT Management Podcast #004 – Cloud-crazy! […]

  2. […] applications. Even if they wanted to, they probably couldn’t. Instead, as we discussed in last week’s IT Management Guys Podcast, some company needs to come along and provide the bridge-code that helps all of those legacy […]