Applying cloud technologies – IT Management & Cloud Podcast #68

John and I spend most of our time recounting SCALE and Tivoli Pulse from last week, leading a discussion about companies large and small having to worry about more connected infrastructure. Also, a couple of acquisitions of note from last week.

Download the episode directly right here, subscribe to the feed in iTunes or other podcatcher to have episodes downloaded automatically, or just click play below to listen to it right here:

Show Notes

• iPad talk
• US/Canada Hockey game
• John goes to SCaLE – infrastructure stuff mixed with Linux passion. Facebook gave a good overview of their setup.
• Facebook: Hadoop, Hive for log scrapping stuff for management. There’s also Nagios and cfengine. Why do they use custom stuff vs. off the shelf?
• Tivoli Pulse – “Integrated Service Management” – mashing together of Rational/TeleLogic and Tivoli/MRO visions – managing everything, even its software. 5,500 people there.
• Beyond the “Obama Check” vision, how was the “grunt” stuff? John says he saw 3-4 private clouds from banks and energy companies running pretty good stuff.
• BoA and JPMC setups for using private cloud – motivated by being able to reclaim resources, self-service.
• John’s two questions: What’s your ration of sys admin to servers? Of those sys admins, how much time do they spend “in the muck” vs. adding “true business value”?
• CA buys 3Tera – one of the long-standing cloud “arms dealers” goes to a member of the Big 4.
• Citrix Online buys Paglo for GoToManage – MSP interest seems on the up; Coté goes over the Citrix GoTo strategy for Paglo.
• NoSQL Boston – March 11th, still a pretty good conference for catching up on new things.
• Coté goes on a vision quest about the internet of things blowing up in the mid-market. Now all we need is easy capital and funding. Arrayent is someone interesting here. Looks like we’ve swallowed this whole IBM Smart Planet vision thing.
• John gets the Scoble experience.
• ITSM Weekly The Podcast – good stuff if you like this kind of thing.

Transcript

(As always, I don’t fully check out and correct transcript for this show, so if we’re saying something nutty below, be sure to check the audio before you think we’ve gone off our meds.)

Michael Coté: Well, hello everybody! It’s the 1st of March, 2010, and this is the IT Management & Cloud Podcast, Episode 68. This is one of your co-hosts, as always, Michael Coté, available at peopleoverprocess.com. I am joined by the other co-host.

John Willis: John Willis at johnmwillis.com.

Michael Coté: Ever since you got on the Mac, johnmwillis.com, I think your audio quality has gone way up.

John Willis: You may be right.

Michael Coté: It’s the premium of quality, that’s what you get with the Mac.

John Willis: I am still trying to figure out how to use this crazy thing.

Michael Coté: Like, after talking about how this is quality, I bet there is going to be something wrong with the recording when I check it afterwards. My voice is going to come through like elephants or something like that. That will be fantastic.

John Willis: Sounds good now though, sounds good.

Michael Coté: Yeah.

John Willis: We might get an A+ rating from the SplunkNinja.

Michael Coté: You know, let’s start out with something completely off-topic for IT Management & Cloud.

John Willis: Sure.

Michael Coté: What do you think about this whole iPad like problem, stink up, what’s going on here? What’s your opinion of the iPod, iPad, whatever the hell it’s called?

John Willis: I think it’s cool. I mean, I don’t know. I would probably get my wife one. I mean, I think — I have talked to a few people and I am not — this isn’t my space of expertise, but I think that what’s cool about it is like, it’s the computer for people who don’t own computers. It’s kind of like the iPhone.

My wife’s brother’s wife, so my sister-in-law, I guess, she hardly ever uses a computer, but she has an iPhone, she loves it. And then she got a Kindle, and I am sure she will be dumping that one.

So I think for books and just accessing web apps and stuff like that.

Michael Coté: Yeah. Well, I had the, I guess, chance is a good way of putting it, I had the chance to basically use my iPhone for most of my work for a couple of days last week, because I didn’t want to take out my laptop. I have talked about this before when we used to talk about Netbooks all the time, about a year ago, and it’s just almost there for like a device you can work on, or someone like me can work on for like two days, where I am not programming or I am not sitting and producing content for like a four-hour stretch, right?

There is a lot of my work that’s basically like checking email, scheduling things, talking with people through email, and doing like short bursty things, if you will. And man, it’s just like so close to the iPhone, it’s so close to doing that. The thing where it breaks down is, if you have to schedule a meeting, like it just is terrible for that, bouncing back between your calendar and stuff like that.

And then also, if you need to — like despite what commercials kind of show you, like there are some good commercials about, we had our iPhone on our family trip and it was perfect. I checked into my airplane on the way and I found a place for the kids to buy candy.

John Willis: Right. And while you are driving at 70 miles an hour.

Michael Coté: That’s right. Like if you need to like book a hotel, it’s actually pretty terrible. Like, if you are lucky enough that the one hotel you want to book at has an iPhone interface, it’s kind of good, but then still, anyways — I think really, for me, like what it gets down to, and this is like the thing on the iPad that makes me suspicious of it, since it’s just a big iPhone, multitasking is what I would really want.

Or even if it wasn’t multitasking, like really fast app switching. Like if I could switch between my calendar and my email as fast as I can switch on my desktop, then most of my complaints would go away.

There is also a lack of spellcheck in applications I find really annoying. But I did find out — I use Evernote quite a lot nowadays, and ever since I got the iPhone 3GS, the newest iPhone; my old iPhone was terrible, it was just too slow at doing most anything. I will just blink and make that statement. But now I have the new one, I use Evernote all the time, it’s great. I highly recommend Evernote.

I would type stuff in there and there wasn’t a way to undo things, which you need a lot, because it’s easy to screw up typing. But then someone reminded me, in Twitter, that you can shake to undo text, which is pretty awesome, I had forgotten completely about that. But that actually — you can shake to undo and then if you shake again, it will offer you the chance to redo it, which I think is great.

But anyhow, the big stink up about the iPhone is basically that — I mean, the iPad, there’s too many Is, there is not like Flash on it and that causes all sorts of Open Web consternation and things like that. Is it a FlashPoint, is that what they call that?

John Willis: Yeah. But watching movies on it would be pretty cool.

Michael Coté: Yeah, definitely. Yeah, like we watched a lot of the Olympics recently on TV, and it would have been nicer to have like — TV is a shitty way to watch video nowadays.

John Willis: I am not there yet. Did you watch the hockey game yesterday?

Michael Coté: No, no, but I was talking with — I always say their name wrong. You might vaguely remember them, they are pretty low profile, these guys, Versiera.

John Willis: The Vegans?

Michael Coté: No, Versiera, and they are a Canadian company that does interesting IT management platform, and I was talking with Frank from Versiera. Anyway, he is a Canadian. I think he is up in Toronto, and he was asking me if I had seen the game. He gave me a recap of it, but that’s the only knowledge I have. I hadn’t had the chance to catch up on it. But how was it?

John Willis: Oh, it was great. I mean, I have been a hockey fan most of my life. I don’t pay as much attention to it these days, but it was just — U.S., which was really not even supposed to be a great team this year, as opposed to Canada and Russia. U.S. beat Canada early on in the qualifying rounds. Canada had to play a bunch of extra games. It was just U.S. meeting Canada in the finals, in Canada, home ice. I mean, it was just — it was —

Michael Coté: Yeah, and then Frank was saying there was like last minute overtime and all sorts of stuff.

John Willis: Yeah, they do sudden death. So they tied. So yeah, so the U.S. came back, got it from, what was it, was it 3-1 or 2-1, I don’t know, but they scored the tying goal like 28 seconds left in the game, which is like amazing. It was one of the best hockey games I have ever seen, and I remember seeing the Lake Placid, U.S. beating Russia.

Michael Coté: Was that the ‘Miracle on Ice’?

John Willis: That was the ‘Miracle on Ice’, yeah.

Michael Coté: The NBC announcers couldn’t go like 30 seconds without mentioning, the ‘Miracle on Ice’.

John Willis: This was pretty darn exciting. It was pretty exciting. So that was hockey. IT Management guy, Podcast fan, hockey fan.

Michael Coté: So also last week, I missed going out to Pulse, I couldn’t make it this year. But you actually went to two shows last week. You went to SCALE, which was in the weekend, and then you also went to Pulse. So I was excited to record this week so you can tell me what happened at both of those conferences. They are both events that I was envious of you being able to go to.

Well, let’s start out with SCALE and then we will get into Pulse. So what is the deal with scale first off?

John Willis: SCALE is what, the Southern California — it’s a Linux — it’s their Linux, but it’s a huge conference. I guess it’s a regional Linux, but it was — they had a four exhibit room — Southern California Linux Expo. But their expo floor was as big as some big damn conferences I have seen. I mean, they had like rows of booths, and like lots of companies.

In fact, it’s funny, because lot of weren’t 8:17 open source, which I thought was kind of funny. But yeah, it was definitely — I mean, it was a great show. I mean, I will definitely go next year.

Michael Coté: So I always wonder, a lot of IT management vendor folks that I know, they end up going to a lot of Linux conferences. Do you think — are Linux conferences about Linux or are they just about like infrastructure nowadays?

John Willis: It’s across the board. I mean, they had — there is a lot of geek Linux stuff, guys trying to jam some type of tool on a box that shouldn’t even be running anymore. I don’t know. There is definitely a lot of weirdness. It’s the classic Linux, people dress up, and some really oddball things you see.

I mean, one of the sessions that was awesome was, Facebook gave pretty much — they walked through their whole stack, open source stack, and what they use each component for, all these tools that’s here. They basically walked through a good hour of what Facebook does, what open source stack tools they use, and then they answered questions at the end too. So that was very cool.

Michael Coté: Any interesting revelations about what they use?

John Willis: No, I forgot all of them; it was cool stuff though.

Michael Coté: Fantastic!

John Willis: So I will leave you up for that. Facebook uses, and I am not going to tell you.

Michael Coté: So as far as like our whole Cloud dev/ops, kind of like crazy automation sort of stuff, was there interest in the audience for that kind of stuff, or was it still kind of leading edge business?

John Willis: There was a Puppet Session and it was pretty crowded. So there is a lot of interest — it’s kind of like OpsCamp in a way. I mean, if you throw up a monitoring or a configuration management session, people will come see it, but it’s just — but there was like Facebook, there was — some guy got sick and I was able to give presentation on Opscode Chef, so that was kind of cool.

I was hanging out with Mark Hinkle from Zenoss, so he had let me come in on their community day to give a presentation. Then he introduced me to, which was the Friday before, but then he introduced me to one of the guys that was running the conference, and I just mentioned, hey, if anybody gets sick, which usually somebody does, and sure enough, somebody couldn’t make it, so I got a Saturday slot, giving a presentation. I got a pretty good crowd.

Michael Coté: Well, that’s 10:49.

John Willis: Yeah. Well, you have got to be ready to run. But Facebook, I was kidding about Facebook. Facebook, it was very cool. They went through like — everybody knows they are a PHP shop, but they had — because of scalability, what they do is they let all their guys, their app developers, develop in PHP still, and they have an open source tool called HipHop that converts it into C++. So they have this stuff around conversion.

And then, they talked also about this tool that I had followed a while back called Scribe. I thought they had canned it, but they use it for — it’s kind of like their Log Scraper program, kind of like Splunk.

But what they do is they feed it into Hadoop. So they had all their Syslog going in through their Scribe to Hadoop, and then they use Hive, which is that SQL abstraction. So it was pretty cool.

Then they talked about how to use memcache. They use memcache like real heavily, which is no surprise.

Michael Coté: So are they using like Hadoop and Hive for management or for the application part of it?

John Willis: They use it for a lot of things. I mean, the one they talked about specifically was for management of log; bring information out of logs to detect, which would be clearly an operational IT management function, but it’s — so they use Scribe/Hadoop/Hive, but apparently they said that a lot of their analytics is Hadoop/Hive across the board, so all the feeds that go into Hadoop and use Hive.

And the thing about Hive is, I had asked the guy who was like managing it, why didn’t they use cascading, which is a really cool abstraction, and he said that they like Hive because its got the — actually, they hadn’t heard of cascading, but the thing is, is that the SQL — the Hive is kind of a 12:44 SQL interface for Hadoop, and it works real well with their analytics guys, because SQL against MySQL; that’s the only thing they run, like a thousand MySQL servers.

Michael Coté: Yeah. I think that’s the — it’s interesting you come across companies like Facebook and they seem to cobble together their own management infrastructure, if you will, rather than necessarily — I don’t always hear a lot about off the shelf IT management stuff being used, like it’s usually a combination of open source stuff.

I am always curious — I don’t know, I mean the assumption — there are two assumptions I always make and I have no idea if they are true.

One is like, like you are getting to, it’s a scale sort of thing, where like the scale wouldn’t work with other systems, but then that doesn’t — I don’t know if that necessarily holds water, because supposedly at enterprise scale, all sorts of systems work. But I think it seems to be more of — I kind of theorize that what they are interested in monitoring isn’t really a standard thing to be monitored, if you will.

So you can’t just go buy, I don’t know, Facebook MySQL log monitoring off of one of the big four, whoever, for monitoring, although I don’t know.

The other assumption I sometimes make is that they just make one to make it on their own rather than get it off the shelf.

John Willis: Well, it’s clear — well, I talked to one of the guys, and, for example, they use Nagios and they use Cfengine, and here they have written their own Log Scraper with Hadoop. So I think part of what you are saying is true, that it’s that kind of Google, the way Google does is completely different than everybody else. And there is no doubt, I am sure Facebook does it different than everybody else.

But there are basics of IT management, and I think that what they wind up doing more than anything is the holy grail, that the enterprise has a hard time getting to, which is taking the tool, getting what it does. And I am sure that, like in the case with Nagios, and they probably had to add some secret sauce to make it scalable for Facebook, but today it’s about getting the raw data and turning it into meaningful data.

So I think that’s their secret source is, they are probably better at — I mean, when you get to the point where you are moving log data, 15:06, you better have an idea of what you are looking for or you are really just wasting your time.

Michael Coté: Yeah, this is one of the interesting things, back at OpsCamp that Mark Cathcart, who is at Dell now, he was kind of — he was asking questions, it was some sort of statements implicit in them but also making statements. So Mark Cathcart used to be at IBM, and he had been there for a long time, and now he is over at Dell and he is basically working on like Cloud scale systems for Dell, as he was saying at OpsCamp.

And he was asking some interesting questions about, what are you interested in monitoring, and how far down does your root cause analysis go, and all these very traditional IT management stuff. And what he was saying is that, in the role he is in, in designing systems for Dell that are used sort of — not that they are used at Cloud scale but they are used for Clouds or for Cloud scale stuff, is that, people want to have really cheap boxes that are very replaceable, and so at some point, at a box level, they are not really even interested in figuring our what’s wrong with the box.

So the kind of traditional metrics of like processor, and he was even kind of joking like fan and temperature and things like that, like people don’t necessarily care about that, they care about some other sorts of metrics.

I mean, he was kind of asking the crowd like, what are those other metrics that you are interested in? What are the things you are interested in? And maybe — and this is all just me theorizing off the top of my head, maybe that’s the reason you see a lot of custom IT management in high scale public website stuff is, they don’t really care about traditional metrics that traditional IT management cares about, they care about these other metrics that aren’t really collected in the traditional way. So that’s why they have to rig up their own Hadoop and Hive infrastructure to get, I don’t know, whatever wacky Facebook — whatever the service management metrics are that Facebook uses to judge, that their service delivery is using, it seems to me is maybe what they are pulling from.

John Willis: Right. But I think — I mean, again, part true on both Mark and/or your assumption, I mean partly, I think that it’s — yeah, clearly nobody cares about — if you are running at scale infrastructure, a fan, I mean, the Cloud doesn’t — there is no fan, right? I mean there is, but it’s not something you are ever going to get.

But even if you are running like Cloud scale Dell equipment, so you are buying pods from Dell, and even in there, a fan is more like the guy who has to replace light bulbs. I mean, so that’s not the guys who are trying to do SOAs, right? I mean, he might have his own SOAs for making sure that not too many of those fans flow.

But I do still think there is a lot of traditional data, even at scale, that becomes important, that does get cleaned into, in the case of, their Scribe, which is an aggregator for Syslog.

And even like, for example, they use Nagios. So in a lot of Nagios, I am sure they are getting a lot of traditional stuff from Nagios, along with their secret source, but they are real — the other thing I asked them about, like for event management, that’s a black hole in open source, there really is no real at scale, something like what IBM has with their Micromuse OMNIbus technology, that’s an at scale event aggregation tool. So that’s what Facebook said, that they wrote their own.

Michael Coté: You should have told them to check out RiverMuse.

John Willis: Yeah, okay, yeah, sure. Well, I mean, that’s a possibility. But I think — yeah. But apparently they found on their own that they — again, there was kind of like, in the open source world there is nothing really, really great for log aggregation and there is nothing great for source aggregation.

So I think to me, again, I think it is — there definitely are some — maybe I am rambling here, but I think there are some type of data points that are unique, but all in all, I think the real magic, if there is magic in these places, is have it aggregate a lot of the common and the detailed stuff together.

Michael Coté: Yeah. Did they talk anything about like the — I am laughing, like trying to pick up, but like the service desk they use or any sort of that other process stuff?

John Willis: That would have been a really good question, yeah. No, at the end I was really trying to pound out some of the ones they didn’t cover, like, okay, what do you do for this? I mean, we have got to go now. Alright, just one last question.

Michael Coté: Yeah. Like on the Agile Executive Podcast, we finally had John Allspaw, who used to be at Flickr, now he works at Etsy, and so we had him as a guest just to talk about dev/op stuff, and to talk about the same stuff he is known for talking about, doing ten releases a day and how the developer in the operations team like get to doing that.

I went back and watched the Velocity presentation he did last year. And it was actually like, I had kind of forgotten how detailed and good that talk is, because it really is, like a lot of the talks, including a lot of rambling on that yours truly does, the sort of like — they sort of lack for actual details of how things are running, it’s more aspirational and whatever, hand-wavy, as James Governor would say.

But in their talk, there is a little bit of that, but it mostly gets to like, here is exactly how we do it, and there is actually like screen shots of the process and the tools and everything that they use.

And they didn’t really get into detail of tooling the process they have, but you can kind of pick out from the different tools they use, the actual sort of like release management tools that they use, and the process that they use and everything. And there really wasn’t any of the — I mean, they were essentially using bug trackers and customized stuff to do their release management, keeping track of the configuration that they are in, and the release, and incoming tickets and stuff like that.

So I mean, that’s another area where, I am trying to figure out the connection of traditional IT management into this new way of running infrastructure. There doesn’t seem to be.

ITSM, or IT Service Management, doesn’t seem to have sort of like thrown enough grapples over this sort of canyon here.

John Willis: It’s interesting that you should say that, because at IBM Pulse last week, in the keynote, that was one of the themes that the Rational guys were talking about, is marrying development and operations together. Now, I didn’t see a whole lot of examples of how they plan to do it, but that was a theme.

It was interesting, because it was really — I mean, it was the same message that we are hearing at the dev/ops in the small place, the dev/ops community or agile operations. So they were saying the same terms, but I think they realized, the world is changing and Rational is kind of that area where — that’s their continuous integration or development process, right?

Michael Coté: Yeah. And that would be the, I think — I actually, while I didn’t get to go to Tivoli Pulse, the analyst people, they are always really nice, they put together like an hour-long preview of all their conferences, so I watched the recording of that. And it sounded like that the phrase they are using is Integrated Service Management, it’s like their theme.

And since you were actually there, I will just lay out like what I gathered from the preview. It sounds like what they have done, and by them I mean — it’s interesting, a lot of IBM is like this nowadays, a lot of IBM software, and to some extent the systems people are like this, but instead of really speaking about Tivoli as a separate entity, a lot of the conversation IBM has nowadays is about rolling up everything, all of the different brands, if you will, into more of a unified message.

So while people like Rational; ever since Rational acquired Telelogic, Rational has really been on this messaging or positioning themselves to sort of like, what would you, how would you put it, sort of like complete management of software throughout all parts of your business, if you will, and tracking how that software goes through.

So Telelogic was a company that basically did — it was kind of like the Rational of the embedded and systems world. So Rational is for places that are purely like software that runs on computers, that does software stuff, as we would know it. Whereas, what Telelogic does was stuff like software that runs in cars, or software that runs in devices. So software is kind of secondary to the hardware that its running in.

So Rational acquired Telelogic, I don’t know, two years ago now. And Rational’s messaging has been a lot more — they always talk about, hey, cars are more complicated than the mainframes and rollerballs and stuff like that. Cars are like, what was that Danny Sabbah said at Pulse, cars are like data centers on wheels or whatever.

So Rational’s messaging is all about software. This is the summary of it, software is everywhere and you have to manage the creation, deployment, management, and then the loop going back on itself. You have got to manage the life cycle of software, and that’s what Rational does.

John Willis: Dude, I am going to summarize this for you. It’s the Obama check. When they bought MRO, and I don’t know if that was before this technology, they wound up getting into a world they never realized that they could do, and it’s basically management of cities and management of — because MRO brought them into a world they never dreamed, which is a kind of physical asset, right?

Michael Coté: Yeah, yeah. Totally.

John Willis: Yeah. So you start combining the fact that you can manage all the assets of a corporation, physical and software, or IT related, and then you can start managing the software development process of everything, including the software that goes into a computer and a car, now you have got big game, now you can —

Michael Coté: Yeah, that’s exactly the confluence of nice purchases and everything. So yeah, like you were saying, on the Tivoli side, there is a similar — MRO is a similar purchase, which bought the Maximo. Now, at least the main thing you hear IBM people talking about, coming out of Tivoli is, Smart Cities and smart this, the majority of that stuff is all the Maximo MRO stuff that gets them into that.

And then, the aspirational thing is like, essentially, if everything is IP addressable or somehow network addressable, then it’s just like managing a data center. You have the same foundation.

Anyway, that was the impression that I got of Pulse was basically the marrying together of that Telelogic Rational messaging and the MRO Tivoli messaging to kind of be like — that hints the Integrated Service Management, but — was I kind of on about what they were talking about?–

John Willis: Yeah, it’s funny — well, first up, at Pulse there is really kind of at least two conferences going on, and then there is maybe three, I don’t know, but there is clearly the whole MRO — they had 5,500 people there this year.

Michael Coté: Yeah.

John Willis: It’s the biggest one they have had in a long time. The MRO Group, they just bring in — there is a lot — they got a huge customer base when they acquired that company. So there is like a whole — like it’s at the MGM Grand and like the second floor is like a different conference. So it’s pretty much MRO, and it’s all that physical asset and it’s that big picture and that kind of stuff. There was probably a lot of Rational presentations, but I didn’t see — maybe not, and most of the ones with MRO is about IT service management, right?

Michael Coté: Yeah.

John Willis: And then you had kind of the Cloud sessions going on, and then you had really just kind of bread and butter Tivoli, you know what I mean? That’s a big part of the conference. So like, yeah, they have this kind of theme message when they start-off the conference about the Big City, we are going to manage cities, and the whole — but then, there is a whole — at least a third of the conference is just focused on the bread and butter Tivoli technology, like monitoring, provisioning, configuration management, and there is really no — I mean, as much as they want to talk about the vision, there is very little connection between the grunts who are just trying to get self-service provisioning of servers and the grunts who are trying to do monitoring on a day-to-day basis and trying to make that all fit together.

Michael Coté: Yeah, yeah. I mean, on the other end of the spectrum, for all the Obama check stuff we were talking about, there are people who are like, hey, how do I manage 5,000 Xen hypervisor instances?

John Willis: Yeah. There is a whole core of people that just come to do that.

Now, the good thing is that the Cloud sessions, like last year when we went, they had just come out with the Tivoli Software Automation Manager, and I thought — I was very impressed as a first cut as what I would call a Cloud broker.

So people make fun of, myself included, periodically about IBM and their Cloud, but I mean the bottom line is, I saw three or four private Clouds from large companies, banks and energy companies, that are running reasonably sophisticated Clouds; now, some better than others. And you could argue that it’s not really a Cloud or there is no such thing as a private Cloud, it doesn’t matter, right? BoA says they are running a private Cloud. If you want to argue with them all day long, you will get no money from them and you will —

Michael Coté: Yeah, yeah. It’s sort of like — it’s not really interesting how definitial they are being, it’s more interesting what they are doing with that. I think now, like increasingly private Cloud means taking advantage of cloud technologies, right?

John Willis: Yeah.

Michael Coté: What that ends up looking like is sort of irrelevant, what’s interesting is that you are doing something new and different, and it may not — anyhow.

John Willis: It’s interesting that, like there was at least — like CenterPoint Energy, CSC, Bank of America, and JPMorgan Chase, all gave presentations on —

Michael Coté: Oh really, they got BoA and JPMorgan?

John Willis: Yeah. So I am saying, last year was a lot about, what is TSAM, and this year there was a lot of really good customer presentations on where they are on the Cloud roadmap and what they have done and what they have accomplished. It was excellent, I mean, to see how the enterprise is dealing with Cloud.

Michael Coté: So since they had some big name enterprises there, I mean, what — this question comes up a lot, and I was actually talking with someone in Twitter about this recently. So what were the justifications for doing private Cloud stuff versus public Cloud? Did they mention that kind of stuff or —

John Willis: Yeah. No, in fact, JPMorgan Chase was a good one for that. In fact, the two best ones, in my opinion, were the Bank of America and JPMorgan Chase.

JPMorgan Chase was interesting like, he listed like the reasons, the motivations for getting into Cloud. What was interesting, the main motivator or the main reason they got into this Cloud thing was for server reclamation.

So everybody thinks about provisioning, they said, I will be honest with you, yeah, they did have provisioning time lapses, with VMware, it was like two weeks, and now they are down to two minutes or could be longer, depending on how it’s handled from a service request standpoint, like is there approval process or what.

But the point is, that wasn’t really — like at the end of the day, that wasn’t killing them. I mean, it may have made business easier for them, but what was killing them is, it just never got servers back, right?

Michael Coté: Yeah, yeah.

John Willis: I had never really thought of that, that was — that we just don’t reclaim services; in this case, servers, but the efficiency they gain by putting a process in place where you do get them back, and it’s the Cloud thing. People with virtualization typically ask for something. Like with Bare Metal, you never get them back. So just moving Bare Metal to Cloud virtualization or — not just virtualization but a Cloud, and what you mean by Cloud is really self-service. You have this ability to – Bare Metal just stays there forever, everybody is afraid to take it back, how do you —

Michael Coté: So does that go into the column of like cost savings, or is it like — I mean, the idea of being able to reclaim servers is great and everything, but like what benefit do they get from that? I mean, to ask a perfectly naive question, but what were the reasons that they wanted to reclaim servers?

John Willis: Well, because you are just wasting resources. I mean, you know that you are over-provisioned. We always talk about being over-provisioned for underutilized servers, but this is a 32:46 where you are completely over-provisioned, because you don’t even know what you have out there that you can get back, and even like virtual instances, without a Cloud-like system.

So really what they have done is they put a front end, self-service interface in front of virtualization and provisioning of virtualization. By putting a schedule in place, it does two things, right? One, it gives people the feeling that after they have tried it a few times that if they give it back, they can actually get it again. You know what I mean? Which is, about the self-server scheduling thing is like, you get it and you say, oh, don’t give it back, you will never get it again, it will take forever to get a new one.

Whereas now, there is this confidence that I can say, well, I need it for like two weeks, maybe it will last three weeks. I am pretty sure that if two months from now I need it again, I can get it, because they have been pretty efficient when I ask for it. That starts like really zoning in on the wasted, whether it’s virtual servers, which really equates to hardware at the end of the day. You get a more efficient use of your whole infrastructure, because now you are actually getting resources back and throwing them back in the pool.

Michael Coté: So were they using like Tivoli stuff to do all that, or did they go over kind of —

John Willis: Yeah, they actually used the Tivoli Software Automation Manager. In fact, somebody asked me, we had a good discussion. I said, that really is a Cloud broker, and they are like, we have been thinking about renaming it. I am like, well, call it Tivoli Cloud Broker, because it really is a Cloud broker.

But JPMorgan Chase actually wrote the original prototype with TSAM, so conjunction with IBM. So they have been working on this project for quite a while. But yeah, they did use TSAM.

The other thing they said I thought was — I mean, this is kind of common, but this kind of falls into the space that I am in, which is that the — one of the things that you have, particularly with development, I heard this at SCALE too, there was a guy giving a presentation on provisioning services.

The thing is, is that image management can be a pain in the butt, right? But we talk about image 35:10 and we think of it from a catalog problem, but actually it’s a development and a bug problem, right? Like the slightest little difference between an image, when it’s out in production, could cause a difference in the way the code works, right?

Michael Coté: Yeah, yeah. I mean, every time you hear of some big dot-com site coming down it’s — not every time, many times when I have been looking at this recently like it’s always some configuration problem, which is kind of funny. And whether that’s inconsistent configuration or a configuration someone sort of didn’t realize was out there doing something wrong.

John Willis: Yeah. And that’s where you get into this kind of idea that — and they weren’t totally up to this place, but they decided that one of their problems is not having consistent images. So I saw this with like all of the guys, the CSC and the BoA, trying to kind of move towards this, just enough operating system and building on the fly. You know what I mean? So that you really do have control.

If you are always building the image the same way, then the likelihood of a configuration option being different — they are not quite there yet, but that’s the theme of what we are doing at Opscode and Luke is doing at Puppet is, infrastructure is code, right?

Michael Coté: Yeah. It’s like the throw it out the window test, right?

John Willis: Right, it’s the throw it out the window test. So they are all realizing that, and it’s funny, but they are kind in the early stages of that, because none of them are really doing — they are still just really taking images, but they are being really strict about the base images and what goes on at this point. You know what I mean? And in most case — in all the cases of like JP and like CSC, I think, most of those guys are still in, this is a QA/dev problem. You know what I mean?

Michael Coté: Yeah. No, I mean they want to make sure it works when they roll it out to production, which is fair. That’s not really too shocking.

John Willis: But BoA was really interesting, because apparently, I got to talk with BoA guys quite a bit, and the thing is that, there are a lot of interesting things happening there, is that, BoA actually has a better vision for Cloud computing than most enterprises I have talked to, because most enterprises still think the public Cloud is the bogeyman, at least from an inside the glasshouse perspective.

We know that a lot of corporations are going out and using Amazon, like there are marketing groups, and people throwing out Facebook apps for 37:48. But the glasshouse is still like, oh, public Cloud, we have got to be careful. With BoA, apparently it looks at the Cloud as a three-tier structure, as private, local infrastructure, or a private internal Cloud, private external Cloud, and then public Cloud.

But what was really cool is, they are actually — they are not going to abandon — it’s hilarious. A lot of companies — and I preach this when I talk to bigger companies, when I get the opportunity is that, embrace the fact that there are people in your company using public Cloud. Most companies ignore it like, okay, well, we don’t have a process for that, we don’t know what to do. We will make believe it’s not happening, and we will go with this, nothing in our company can go outside our firewall, but meanwhile there is –.

BoA is taking a realistic approach and said, alright, you know what, we are going to embrace everything and we are going to apply; you would have loved this presentation, so Guy was basically their service management guru, their top kind of idol guy. He went ahead and he mapped, for this three-tier structure, they are still evolving, but he mapped all their service management pillars, their iPhone structure or process that they have defined for how it’s going to work with all three types of Clouds.

Michael Coté: Right, right, right, that makes sense. So it is a mapping of like existing and legacy to future, if you will, but it’s Cloud stuff.

John Willis: Yeah. Well, what was funny is that, when the first group started doing Cloud stuff in there — in fact, what happened was, I guess Merrill Lynch was a lot further along. They had done a lot of research on Cloud and all that stuff, and so when they acquired Merrill Lynch, they actually got a lot of intellectual property on how to do Cloud from Merrill Lynch.

But the thing was that — so they were first coming in they were like, the service management guys, the monitoring, all the core guys that managed everything, very sophisticated; I have done a lot of work for BoA, so I have been inside their glasshouse, but the Cloud guys are like, well, you don’t need to worry about this stuff. And we are like, well, we think we would like to worry about it. They are like, no, no, it’s cloud stuff, don’t worry about it. You don’t really need to be involved.

And the service management people were like, baloney. And they sat in on all the meetings, and as they got further along, they actually comminished 40:24 the Cloud guys, like this is IT management 101. There is no pixie dust in the Cloud.

And that’s the beauty. That’s what I thought was great about their story is that, even not me, when the Cloud first came out, I said, it doesn’t really matter in the Cloud, and I was so naive. Of course, process matters.

Michael Coté: Yeah, yeah. I mean, that’s the cutesy thing I always try to harp on about Cloud stuff is — and more generally, this seems to apply to lots of, what would you call it, technology process innovation, is that, you are actually applying a lot more discipline. It’s not necessarily more discipline, but you are applying better discipline than you had in the past.

And to the point of like a lot of Cloud stuff, I mean there is like — you are trying to operate at such scale and like at such dynamism, if you will, that if you don’t have discipline, like things can go really wrong.

I mean to be analogous, it’s like, since we were talking about the Olympics, I was watching all the bobsledding stuff, and the guy is like, basically, when you are going 80 miles an hour down this track, there is no room for air. Like you scrape the corner of your thing against the sidewall and like you are done, like your thing flips over and that’s it.

And it seems like that’s the thing with a lot of — like if you are going to set up something to automatically provision like 5,000 servers and do all this stuff, it’s like, there is not really a lot of room to like cowboy code it, or to like cowboy it. So you do need to have this discipline of like, when we do this button, it’s really going to work. And rolling it back is not going to be — it might be possible, but it’s not going to be pleasurable.

So I mean, I think to your point, exactly, like of course there is going to be a lot of process involved in it. I mean, there is going to be — as that old famous quote goes, a lot of effort went into making this effortless. And I think that’s what people like forget about it.

And analogously in the tech world, like agile software development is like this, where there is actually — it may look like there is not a lot of process and discipline going on with agile, but there actually is like a ton of discipline on an hour by hour, day to day level that people are doing and that kind of stuff.

John Willis: Yeah. It was funny because both sides learned from each other. So the service — the kind of brick and mortar, we have got process guys who are like, no. The Cloud guys are saying, no, you don’t worry, you don’t have to worry. And they are like, maybe — they were like questioning whether they really were being kind of onerous, or, oh, there is those guys again. But they said, okay, let’s just stick around. Let’s just go to all the meetings and before long —

Michael Coté: Yeah. I mean, it’s funny, like you could see on both sides, like you could have the ITIL guys and they are like, well, if we are going to do release management, we need to draw up a manifest for the configuration and all the changes that are going to happen. And then on the Cloud side the people are like, oh my God, that’s so much work. We don’t have time for that and everything. What we really need to do is draw up a model of how everything is going to be done and then specify declaratively how this model is automated.

And you realize they are talking about the same exact thing. They just like — it’s just like, kind of the way they represent it to humans is like — I am oversimplifying, but the implementation is what’s different. But the front end to it, where people interact, process-wise, ends up being the same.

And there are several like Run Book or RBA and companies out there who are kind of like fast adapting to sort of Cloud ways of doing stuff. One of them that I talk to on and off is this company, newScale, who — I have only talked to them, and I have some of their customers lined up, I think, eventually to talk with, to validate it. Like I have talked with them a couple of times over the years and they have an interesting — they have kind of a very interesting realize that the RBA stuff they have, which is very classic ITSM stuff, like it fits really well, or it can be made to fit well with sort of Cloud conceptions of doing things.

And exactly to the point, it’s interesting, because they didn’t quite — what they were talking about wasn’t articulated in the way of freeing up capacity and freeing up things. But at the end of the day it became Cloud stuff for their customers, they were telling me, it became a matter of — it wasn’t so much like Cloud stuff as we would know it as like scalability and flexibility, but it was more like just a better way to manage all that crap that you have, whether it’s freeing up capacity or provisioning things faster or whatever, it was just a better way to do — it was a better implementation of service management, if you will.

I mean, that’s at the moment what it looks like private Cloud is going to look like, is just like, it’s the same old service management stuff that we are just too embarrassed to call by the same old words, because we don’t want to scare you off, but it’s just a better way of doing it than all the old way that we had of doing it. It’s a better implementation, essentially.

John Willis: Yeah, there was a —

Michael Coté: Is what becomes exciting for — that version of private Cloud, I am a lot less cynical about it than whatever else stuff I get all uppity about.

John Willis: Yeah. I mean, I — you know it’s funny, because I am — and obviously anybody who listens to me rant for the last couple of years realizes that I learned stuff and stuff that I rant at about six or eight months from now, I don’t rant as much. I mean, I used to be pretty harsh on saying that a self-service provisioning system for virtualized servers is a Cloud, because it doesn’t — it’s not like the 800-pound gorilla, which is Amazon, which is more than self-service. It’s an abstraction of hypervisors.

But at the end of the day, I mean who cares, right? It’s what you really need to get out of it. And for these companies right now — that was another thing. So we looked at BoA, one of the messages is that, we can — let me step back.

So before I went to SCALE, I had an opportunity to visit Shopzilla. I had done a podcast with their CIO, and so the day before SCALE, I was able to go out and visit their site. And they wrote their own private Cloud, but one guy wrote it in two months. It has more features than any of the open source private Clouds that I have seen before, right?

Michael Coté: Yeah.

John Willis: It has a whole system for not only scheduling, but it does like service reaping. So like, if you have like over allocated your memory and it notices that you haven’t used half that memory, it will actually kind of help you back. If you keep a service active and it’s not active for a certain amount of time, they will send you an email saying, hey, you have got like 24 hours to deal with this. So all this, I am like, this is the pretty cool Cloud. You know what I mean? They did it themselves, one person, they use OpenVZ. So it was all on one platform, so 90% of what they run, this company runs on CentOS, right?

Michael Coté: Yeah.

John Willis: And it’s OpenVZ, which is the open source container base Cloud or virtualization. So it was easy to build this thing, and they were using OpenVZ already, so they had a lot of their own intellectual property on it.

But the thing is like, so you go to BoA and you look at Shopzilla, it’s kind of this 47:35 and it’s complicated, and they have got a lot of stuff going on there, but the thing is, is 90% of their stuff runs on CentOS, right?

Well, BoA, I mean, gees, their stuff runs all over the freaking place, right?

Michael Coté: Right.

John Willis: So they can’t just like go, oh, just go ahead and throw it all on one abstraction of hypervisors, go ahead and put it on Eucalyptus or Ubuntu Enterprise Cloud, they have got to figure out how to get WebSphere to run on the Cloud. They have got to figure out how to get all these classic legacy applications. So they have to crawl to build their Cloud.

Michael Coté: Yeah. I think it was October last year, Luke Kanies of the Puppet fame and I talked with the Shopzilla guys about their use of Puppet, it was Abe Ingersoll, I think was his name. But I mean, they had an interesting model of — I mean, obviously Shopzilla — I forget how many data centers they had across the globe, but they had really gotten to that point of programmable infrastructure at least for the images that they made and everything, which was interesting to hear about.

John Willis: Yeah. I spent almost a whole day with them, and it was great. I got to talk to — I am not sure, I talked to the CIO for a while, and then I talked to the guy who ran all operations. So one of the things I have been doing these days is asking people two simple questions. One, what’s your ratio of sysadmin to servers? And then of your sysadmins, how much time do they spend in the muck, versus how much time do they add true business value, kind of top line ROI to the company.

It’s interesting, I would have thought a lot of people would you give you bullshit answers or not want to answer it, but the last four or five companies I have asked that question, I have gotten straight up, in your face answers, and Shopzilla was like 100-200 49:38 and about 70% in the muck, so that’s kind of cool.

Michael Coté: Well, so speaking of Cloud stuff, I think this was last week, it was last week, the CA bought 3tera to do some of their Cloud stuff. So that was pretty exciting.

John Willis: Yeah, I totally missed that. I felt like a dote when you told me that.

Michael Coté: I have been working on a little writeup of it, so I don’t really have much of a writeup. I mean, my sort of quick analysis is, like we were joking, they bought the — they basically bought 50:15 or the assets of it, which was kind of — it seemed to be — it was sort of a pre-Cloud, Cloud infrastructure for doing the things, which is interesting, and 3tera is one of the longest — I mean, 3tera was doing Cloud stuff before it was widely called Cloud things essentially. I don’t really have a customer account for 3tera in front of me, but I mean, they seem to be a successful — they are basically a Cloud infrastructure arms dealer.

If you wanted to run a Cloud, they would sell you the software that would do that, and they seem to — I mean I — it seems like they were doing alright as far as the technology and sort of the customer base.

John Willis: Yeah. Well, 3tera, I have done — I have been promoting those guys, and I actually did some consulting for them, but they are actually — the one thing they have got is, they are not open source; I don’t know that that’s the end of the world, but their biggest sweet spot has been hosting company. So they have done pretty well with hosting companies, because it’s kind of, get your Cloud in the box. So they have had a lot of success. And then they had some big deals with some, I think some companies in Japan and stuff like that, so that probably got them on CA’s radar.

Michael Coté: Yeah. You know, a lot of acquisitions that have been happening recently have been — there is a MSP or a service provider tie in, which is of an interest in the IT management space.

I don’t really know the managed service provider or hosters, I don’t know that space extremely well, like I sort of know the vendor side, but it seems like lots of people in the IT management space, that’s where they are getting cash from and getting notice from.

There is another — also Citrix Online, which is a part of Citrix, they bought Paglo last week, and part of that was — part of the mention there was that there was a lot of MSP sort of usage of Paglo, which is one of the only SaaS based IT management platforms for monitoring out there. Tivoli Live is out there at the moment, but there is not a whole lot of other ones.

ManageEngine, the Zoho people, they have one in beta for doing SaaS based stuff, and I am sure there are some I am forgetting here and there. But anyways, they were bought by Citrix, which is interesting.

But the whole point, getting back to it, it’s interesting that MSPs are getting a lot of action out there.

John Willis: Yeah, I wonder though, on Paglo, so I checked those guys out, like I was comparing them to Splunk, I would put them close to the Splunk category, although they — Splunk says they can monitor your stuff too, but at the end of the day they are log scrappers.

But I just wonder if just that space just — they can’t get — I don’t know, I am just guessing, maybe they can’t get a second round, that the only way out of that mess is to get acquired on the cheap by somebody.

Michael Coté: Basically, there is GoToMeeting, it’s the GoTo Group, there is GoToAssist and GoTo all the stuff, so that part of Citrix, that’s what Citrix Online is, and that’s the part that bought Paglo. So it’s not like Citrix bought Paglo, so to speak.

John Willis: Right.

Michael Coté: And I think for them, I mean I think it —

John Willis: What would they do with it?

Michael Coté: Well, I think the combination there kind of makes sense, because what — according to what they have told me, to use that standard disclaimer. I mean, Citrix Online actually has had pretty good growth and revenue and usage and everything, and so what they found is in the small and mid-market. So there is a lot of people who use GoToAssist, which is basically like something is going wrong with your machine, so you want a remote login to like look at it or manage it.

John Willis: Okay.

Michael Coté: So I think with Paglo, there is a lot of opportunity for like these types of like admin people who are working in the mid-market or they might be independent, like if they could sell them Paglo as well, for more thorough monitoring, kind of like the Spiceworks crowd essentially, right?

John Willis: Right.

Michael Coté: Like these people would respond really well to having Paglo as the theory. So that’s why Paglo is being renamed to GoToManage essentially.

So the idea is, Citrix Online already has all these people who are using GoToAssist or GoToMeeting. They have got a lot of mid-market attention, right? And so they are basically buying another product to sell into that market that they have. And then it’s the same — what they are getting from Paglo is — Citrix Online doesn’t really do — they don’t really have the — they are buying the expertise of doing that kind of management.

John Willis: Right.

Michael Coté: The other thing is like the whole SaaS model works out for the way that Citrix Online wants to deliver their stuff, right? I mean, they don’t want to like having an on-premise thing to manage, at most, you have the little Paglo agent that you install.

John Willis: Right.

Michael Coté: So that’s sort of like the plan for the acquisition there. Like I was saying, I think — that’s why I was going off on this whole MSP thing is because — this is a similar thing that happened when the little 55:10 four word targeted more on the mid-market than kind of going after enterprise stuff or the former little four 55:15 or whatever you want to call them, is that, it seems like there is more of a pull from — at the moment, there is this — I don’t know what you would call it, I want to call it democratization, but it’s kind of like bringing down enterprise-grade management to the mid-market, if you will, and it seems like kind of a market that previously wasn’t served too much by higher grade stuff. I don’t know.

John Willis: Well, that’s the thing. Actually, I talked to the Tivoli guys a little bit about the — I mean, the thing right now is, the mid-market is, what we class — SMB is growing like there’s no tomorrow, the kind of Cambrian Explosion, as I have described it for, just infrastructure, everybody can infrastructure now. So you are just seeing a lot more — actually, I was describing this the other day to somebody, I was telling about — explaining to them about the book, ‘The Long Tail’, because they were asking like, so why — it was actually kind of why I was so excited about our market right now, and I said, because — I said if you think about — you get ‘The Long Tail’ concept, that it was — everybody who listens to this podcast surely would get ‘The Long Tail’, right?

Michael Coté: Sure.

John Willis: We had a little cliff, but now we turned it into a long tail, in terms of consumable items. But what I was explaining is why I think infrastructure is like red hot, like dealing with — helping people become scalable for IT management infrastructure right now — because you need to take the long tail, what you really have now is all these spikes of people who never had an opportunity, now have an opportunity to not only become consumable, but they actually have the opportunity to become huge.

In other words, think of the scale now, there’s a lot of spikes in that long tail, in that, these people would never get — some guy who makes Brazilian flutes with purple and green polka dots now has a following, because people around the world can find him.

Well, it may turn out that, that might go viral, whereas nobody would have ever heard of him without ‘The Long Tail’, now ‘The Long Tail’ becomes like individual spikes of like — like Etsy is a good example. You were talking about John Allspaw, right?

Michael Coté: Yeah.

John Willis: Great example of like a really clever idea that would have never got off the ground before kind of the Internet Long Tail consumable availability and then combine the ability to have Cloud like resources. So I think that’s why we are seeing —

Michael Coté: Yeah. I mean, it’s the mid-market SMB vision of all the IBM vision we were talking about earlier, where basically like, if computers are everywhere and everyone is connected, then in that same way, like every small and mid company has a bunch of infrastructure they want to manage, and so they need the tools that help them manage that infrastructure.

So if you want to be a little arts and crafts store and you have to somehow — that’s a bad example, because you just set up a Etsy thing, but whatever it may be.

There’s another example to like really go far out. So I talked with, I guess you would call them an Internet of Things company called Arrayent, I think is how you say their name. What they basically have is — so they have this little device that you hook up to, and by you I mean like a consumer. So you hook up this little device to your home Internet and it creates this like sort of, I forget what frequency or protocol they use, but they have their own sort of wireless network that, that creates for you.

And then they have these chips that will use that wireless network, so you can have — so consumer goods companies can make embedded devices that are Internet connected.

So as an example, like the first thing they worked on was this toy with Mattel that was basically like — it was basically just like a little text messaging thing that you would give to like your, I don’t know, 12-year-old or whatever, some category I know nothing about.

So instead of like giving them a cell phone that they could text message with, they would text message with this little private device network. And from an open source, like Open Web perspective, you are like, oh my God, it’s this private network and you have got to buy a device, that’s terrible, blah, blah, blah, but it gets away from the whole point of it.

The wider point was like, so never mind the toy, it was like, you can hook up your thermostat to this, and just imagine all the devices that you have, and you would buy into this Arrayent system to basically have all your devices on the Internet and connected.

Like their is the iPhone ad of like — it’s the same one, like I printed out the boarding pass and then I found a place to eat hot dogs. And the other one was like, did we forget to turn the lights off, and then there was an app that goes to like turn apps off. And so that sort of vision of having all your shit on the Internet, essentially, to curse for the second time in this lovely podcast.

I think it starts to pull a need to have more — in the mid-market, you need more management of that infrastructure. So as an example, like that’s part of what these Arrayent people do is, part of their sales pitch is like, hey, you are Mattel or whatever like layer in the global supply chain that actually makes what Mattel sells, and you have no idea really how to program Internet scale applications or even pseudo Internet scale applications, and we kind of handle all that to you.

All you have to worry about is programming to this chip and then maybe programming your software on the back end that works with this, but we will handle the scaling up and the routing of the traffic and stuff like that, whether that’s people on the way to the airport turning off their lights or people text messaging or whatever you can dream up. I mean, that is — I think that does mean that there is a lot more people who are relying on technology beyond like static web pages or Etsy sites, where they are selling stuff.

John Willis: Yeah, yeah, just across the board, yeah.

Michael Coté: And I guess being even more macro about it, if that’s the right way, what everyone’s hope is, is that now — the hope is I guess that we are sort of out of these economic doldrums, because all of this vision, it needs lot of cash. So I guess the hope is that, all of these mid-market and above companies can actually start getting loans to actually like come up with these innovations and then also will have the money to actually deploy them and run them and stuff like that.

So if capital is freed up, and we have new technology, where you could actually have all these connected things, then you have got all this crazy stuff happening. I guess that is the one dark thing is, until it’s easy for businesses to go get the loans they need to like build out these — everyone being able to turn off their lights remotely, then nothing is going to happen, but the technology is just waiting for the capital to come alive.

So there is my rosy-eyed view of the future there. I need to go give some talk somewhere, where I can run offstage quickly before people start asking smart questions.

John Willis: Yeah. Bottom line is, there is going to be a lot of new data centers, whether it be virtual cloud or whatever, that didn’t exist before, because of all these things. I describe to people what we do is, like there is — make it real simple, there is three parts to this. There is the data center infrastructure. Well, we have solved that problem. It’s VMware or it’s Amazon or it’s Rackspace. You can get resources, server resources pretty quick, pretty easy now.

Then on the other end of this three piece stick, I guess you want to call it, is the app development. Well, that is what it is. There are some kind of pushbutton built in app things, but at the end of the day people have to write code.

Then there is the piece in the middle, which is, how do you build the infrastructure, right? And we haven’t — I mean the bottom line is that, a lot of people think when you get a Cloud, that’s the pixie dust part, but the truth of the matter is, there has got to be glue to build and manage infrastructure, whether it’s monitoring, configuration management, event correlation, log scraping, log aggregation, those are all things that have to — isn’t it a great way to tie it all the way back from the Facebook discussion —

Michael Coté: We have swallowed that IBM vision hook, line, and sinker man, we are smart planetized.

John Willis: That’s right, yeah. But IBM will put an abstraction layer on top of all of this to manage it all.

Michael Coté: That’s right, exactly. So there is one other like fun thing before we wrap up, but I saw that there is a NoSQL Boston Conference, March 11th that —

John Willis: I have heard of that, yeah.

Michael Coté: I mean, I think those are — as I have told people, I don’t think that NoSQL Conferences have gone kind of sourly yet, so they are still a fun place to go to, to kind of see what’s happening in a new area of technology.

John Willis: So you are going to go?

Michael Coté: No. I have kind of limited travel at the moment, but isn’t March 11th like next week too? I will see, maybe Stephen O’Grady will go there, since he is in the Boston area.

But the next thing I am going to go to, speaking of off topic, is I am going to go to Microsoft’s MIX Conference, which is all about — it’s kind of like the front end for all of this, this blue sky we have been talking about John, like Silverlight and IE and HTML 5, it’s all the platform and UI layer stuff that Microsoft does. I will be going to that.

And then I should be able to go to South by Southwest here in Austin, which, I don’t know what’s up with these Microsoft people, they schedule — MIX and South by Southwest are kind of like the same crowd, and they schedule MIX right at the tail end of South by Southwest, which is tragic. So bad scheduling. But that should actually be a fun show.

But other than that, I have got — there’s actually like — I will have to put a link to this now, but I have a keynote about a lot of this Cloud stuff we have been talking about at the — it’s like the Enterprise Emerging Technology or an Emerging Tech Enterprise Technology Conference in Philadelphia like the second week of April, and I am giving one of the keynotes there about just what all this Cloud hoopla means for people, so that will be exciting.

But the thing I was going to ask you about is, you recently did a video with the building43 people, otherwise known as Robert Scoble. What was that like?

John Willis: It was pretty cool. So the funny thing is that, I knew a couple of guys from the Rackspace because they were at the OpsCamp Conference down in Austin. So I saw that Scoble was — he was in town for some startup conference they were having and Jungle Disk, because Jungle Disk was acquired by Rackspace. So they were like the Premium or Gold or Platinum sponsor for this startup thing.

So I saw that Scoble was going to be in town and they were going to have a Tweetup; I was like, oh man! We had a customer — I had a customer that did a presentation on how they had been using our product the week before. I posted some of that, how they use Chef, and it was really cool, and I was thinking, man, won’t it be cool if I could get Scoble to interview me and my customer?

So I went to this Tweetup thing and Scoble was there. It’s surreal. I mean, there were just people waiting in line. He is like the Beatles, I mean people just want to touch his shirt, and like, I touched Robert’s shirt, running out of the room. And I am like, oh, this is going to be a nightmare.

And I told the customer, I said to the Rackspace guys I knew, I said, hey, you know, is there any way — oh, that was the other thing, the customer was using Rackspace with the Cloud. So I was saying, hey, that’s what you are supposed to blog about. Went to Scoble, he is like, yeah, yeah, yeah, and yeah, yeah, yeah, call me tomorrow. Yeah, right?

Then this producer was over there and this producer was like, it ain’t going to happen pal, you ain’t going to. So I was with the customer, visited some bar with the customer, I said, it ain’t going to happen, it was worth a shot.

I had my Cassandra shirt on from the OpsCamp that the Rackspace guys were giving out. And all of a sudden some guy walks by and he says, nice shirt. And I turned to him and I said — he works at Rackspace. So I tell him about — that we got this — I got a customer, he is using the Rackspace Cloud and he is using our product. And he is like, hold on a second, Scoble works for me.

So he pulls Robert Scoble over to our table and he is sitting there, and they got a scheduling to do a video the next day. What was interesting about Scoble, and I will tell you about the video in a second, but when I sat — have you ever met him in person?

Michael Coté: Yeah, I met him briefly at an Adobe event.

John Willis: He started talking — I asked him questions about how does he do all this aggregation and all that. I mean, it was fascinating. You started getting a feel for why he is Scoble, you know what I mean?

Michael Coté: Yeah.

John Willis: Because he said that he has like a spectrum of like 4,000 people that he basically knows, you know what I mean? He couldn’t tell them by name or anything, but it’s like he knows them on Twitter and he knows like, if they are going to say something, he knows, if it’s their spin-on, because where they come from — and he was explaining how he kind of aggregates this like large community of Twitter followers. And he is able to use that to kind of being 68:59 knowledge or kind of an Oracle, for things going on.

What was interesting also, we started talking about the iPad, and he was talking about how he envisioned using it for his kind of network of aggregating information; it was pretty cool to hear him talk about how he would maybe have kind of a multidimensional visual on the iPad that you could kind of spin around in circles, so you could have like Twitter lists and Facebook lists, almost like the Tom Cruise movie, where you can drag him in between.

Michael Coté: There’s a video from a couple of years ago where someone sat down and had him show how to use Google Reader, which is probably out of date now. But it was interesting to see, it’s sort of high stream —

John Willis: How he could do that. But yeah, so the next day they scheduled us, we got like the last time slot, it was cool. It’s kind of — it’s interesting, he is just really kind of down to earth and we got a great interview in.

So in fact, I did a bunch — they make you work for it though, because then you’ve got to do a bunch of your own. They make you do the editing, I guess, but I guess it’s kind of —

Michael Coté: Oh! Really? What platform do they use to make it let you do your editing?

John Willis: Well, it’s probably less complicated for most people, but first they wanted me to – well, so they do the video and then they had like three cameras on; one on me, one on — and then they had one camera on all of us. So I don’t know. I guess that’s four cameras. So he sends you the four-camera version and then there is everybody on it and then you are supposed to give him a video back with where you would want to put screenshots and stuff like that interjected.

So you almost have — it’s kind of build your own overlay video that you sent back to them where you kind of timed the way you want. Maybe when I am talking, I will start the conversation then I will show a screen of Opscode website.

Michael Coté: Alright! And that’s what you have to insert in?

John Willis: Right, but I don’t do a lot. So the producer guy was pretty nice. He said that if I sent him like a bunch of pictures, then he would insert them if I told him where to insert them.

Michael Coté: Yeah. Well, you will have to tell people what the benefits and success are for you and Opscode after doing that, because it’s always — here we are in a podcast ourselves and I think it’s sort of uncharted territory as far as like — other than like raw page views like how that translates into something that’s beneficial. I am always curious about what —

John Willis: Yeah. I will tell you it was fun, because if anything, I just got motivated then I was just really kind of spot on, on my answers and also there was a lot of fun for me just to watch it, we did it. So, yeah, it was really cool.

Michael Coté: Definitely. Well, I think that’s about it for this episode, John. Is there anything we left out?

John Willis: No, I think that’s it.

Michael Coté: I agree. Well, if there is anyone coming to South By Southwest in a few weeks, be sure to drop me a line, because I am of course here in Austin. It’s always nice to meet audience members if you will or parts of the community to use the 21st century way of phrasing it.

John Willis: Oh! You know, by the way, a lot of people listen to us. So that is kind of interesting. I ran into a lot of people that —

Michael Coté: Oh! Yeah?

John Willis: Yeah, so we have got a pretty good following of the people who go to 12:43.

Michael Coté: Well, we should use our little black shine squeeze boxes to record some of the little interviews with them.

John Willis: Yeah, yeah, yeah.

Michael Coté: What are you going to do?

John Willis: What are you going to do?

Michael Coté: I am always saying I should do that and I never do it. So who am I to say that it should be done? I am shooting everywhere.

John Willis: Yeah, shooting.

Michael Coté: Alright. Well, I guess we will see everyone next time and thanks as always for listening and it’s always — as John was alluding to it, it’s always nice to hear from people, that’s always a good thing.

John Willis: Yeah.

Michael Coté: Well, I know one thing I wanted to say. I forgot if I mentioned this last time, but there is a new podcast from the ServiceSphere guy from Chris Dancy and he talks — he has got a people in his four episodes now and he talks a lot about ITSM and stuff. Actually, I’ve heard the first one and I will put a link to it. It’s good stuff. If you like the good old fashioned IT management stuff, we talk about here; you will probably like that podcast. So you should definitely check it out.

It’s like this guy does, his name is ServiceSphere on Twitter and he basically does — it’s hard to describe exactly what he does. I think he does like Service Desk Consulting and also helps people run service desk stuff and ITSM stuff and I am babbling at this point, obviously.

But what’s interesting is at least in the first episode I have heard so far, he has got two people who are short of in the industry and they do work and it’s funny to hear vendor side, consultant side, and then industry side people talk about stuff and it’s good stuff, good material.

John Willis: Cool!

Michael Coté: So with that we will see everyone next time.

John Willis: Alrighty.

Disclosure: IBM is a client, see the RedMonk client list for other clients mentioned or related.