The itmanagementguys Tag
First, we throw out a little pro-top for you: if you want to see a good idea of what we’ll talk about each week, check out the del.icio.us tag “itmanagementguys”. Also, if you want to stick something on our radar, feel free to tag it with itmanagementguys yourself and we’ll consider it.
Acquia, Drupal, and CMS
John then jumps into a post from Acquia’s Jeff Whatcott on the continuing quest of Acquia to shake off the category of CMS from Drupal. On the topic of open source, I mention this year’s funny statement from Gartner on open source: by 2012, 80% of software will use open source. As I summarize, if you’re not using open source by now, make sure your boss doesn’t know ’cause you should be fired.
We get back to the topic of Drupal being a CMS or not. In summary, there is some truth to the idea that platforms like Drupal, django, and Apache Cocoon aren’t “CMS” systems in the classic sense, but are used for largely the same goals: making public web sites, if not with richer “write” abilities.
What up with VCs Hating Services?
And then we launch into an extended discussion of the VC/Open Source services paradox/lie. As I note, I’ve discussed this idea with several people. What is it? In summary, the train of thought is that telling a VC that you’re going to base your business on providing services and support is VC poison. If you’re an open source company, VCs don’t want your money-pipe to be based on humans doing things.
And yet! It seems that most cash-outs of VC funded open source companies are based primarily on the acquiring company wanting to provide services around the open source project. For example, Sun buying MySQL for $1 billion dollars. Not to mention the billions of dollars of revenue that companies like IBM, Sun, Accenture, and others make from services. More pertinent to the topic at hand, a huge amount of IT Management sales are around services: getting a legal copy of the software is a tiny part of the monetized elements up getting an IT Management suite up and running.
In short: it doesn’t seem to make sense that services, which can generate billions in revenue, are VC poison. The paradox/lie here, though, is that most open source companies base their business models on services and support. What exactly do VCs think is going on otherwise? But what do we know?
Applications in the Cloud
Next, we jump into the “what and how are applications going to be run in the cloud” bucket on the topic of Google Team Edition. Team Edition allows people to create Google Apps installs around email addresses instead of domains, like how you identify your company and school affiliations in Facebook. I get all excited about this being a way to subvert IT, but John reminds me that he who controls the network controls the cloud: the IT department can just block access to google.com or whatever. Darn!
John’s Cloud Layers
With our foots in the door of our cloud-crazy talk, we jump into the cloud portion of our show. John mentions and we discuss his 0-3 level analysis of what cloud computing is. Without summarize the entire post here (just go read it!) the point is that there’s virtualization at the bottom, some smattering of “grid” through-out, then a total un-caring of your IT’s physical layout – topology even! – towards the top, with a very service oriented (vs. systems/software) perspective on things at the top. Or “SaaS” as we used to call it a scant 60-90 days ago.
At the high level of cloud computing – a new way of running IT – I add more color to the tragic tale of the “Little 4″ phrase, namely why I put openQRM/Qlusters in there originally. Part of my thinking was that the openQRM method of managing IT seemed new at the time and thus, while not a platform for IT Management as we know it, it was something new and different at a platform layer. But, as I admit, I was probably also just charmed by whurley.
We then get into a discussion of how provisioning plays into managing a cloud/grid. John recalls promises of cloud-by-provisioning from Tivoli years ago. Then I launch into a painful metaphor of pass-by-value and pass-by-reference for two different ways of managing clouds: sort of provisioning vs. federation. Even I have no idea what I was talking about.
John then tells us about checking out Cassett in reference to his cloud-craziness. Both them and 3Tera, he says, promise a way to deliver “utility computing.” They both seem to be promising the same results. While John doesn’t quite know which one is better, or whatever, his guy tells him that more closed systems – perhaps the pass-by-reference metaphor – would work better as there’s less moving parts.
At this point, I segway into “the Parade section of the newspaper” and bring up some Cray talk I had recently. First, I was astonished that there were still Crays around – who uses the term “super-computer” anymore, right? Second, I learned that Crays being built today are massive: several sizes bigger than my house. Just one computer, bigger than my house! And, there is no more sexy phrase in systems than “vector processing.”
John then tells us about his time at Exxon where they purchased the first Cray for commercial use. Apparently, there was no I/O system on the Cray, so Exxon couldn’t load up all of their seismic data from tape. He recalls a call with Cray, including Seymour Cray, where the Cray folks were confused as to why they wanted to load data from tapes.
Hadoop, MapReduce, and IT/Business Alignment
What with super-computers out of the way, I admit to John that I have no idea why he cares about all this Hadoop stuff. How does this apply to IT Management?
The summary is that all the fancy greek talk of Hadoop reduces – HELLO! – down to being able to quickly search over massive data sets. What’s important here is that previously un-usuable information streams are (potentially) usable if you “index” them with Hadoop. The point here is that IT can provide the business side of the house with new sources of information to make decisions: “how are our sales doing for widget X world-wide, right now?” and so on.
John tells us about how Rackspace is using Hadoop to look over mega-sized mail-logs. The connection here is to start thinking about how this stuff gets you new ways of doing IT Management, if not the holy grail of “IT/Business Alignment” – never mind the blinky lights, give me the data!
When it comes to being able to do something with stupid amounts of data, more high-powered, direct advertising comes to mind. For example, as I say, most of the reason money-hogs are interested in Facebook is the huge amount of data about people available. People are expressing interests and passions, group affiliations and friendships. For a money-hog, this means an easier way to find reasons to get cash from these people. Do you like summer sausage? Does your boyfriend like summer sausage? Why not buy some! All that Gillmorian Attention/Gesture stuff just might be onto something, if, you know, spam-y.
Getting towards the end of the show, I give a review of Sun’s openxVM platform and strategy. I was most recently PowerPoint-stuffed at the Sun Analyst Event last week, so I brain-dump on the Sun IT Management plans.
Whatever Happened to Chargebacks?
We round out the show with another arcane topic from the past: chargebacks. Chargebacks are essentially internal billing between the IT department and “the business.” As I note, it seems like charge-backs aren’t as precise as they used to be probably, as John points out, because we no longer have the accounting ease that centralized mainframes bring.
Disclaimer: Sun is a client, as are LogLogic and Prism. See the RedMonk client list for other clients mentioned above and in the podcast.