IT Analyst Dan Kusnetzky Talks about Cloud Computing and Cloud Hype (Video)

Video no longer available.

Dan Kusnetzky and I started out talking about cloud computing; what it is and isn't, how "cloud" is often more of a marketing term than a technical one, and then gradually drifted to the topic of how IT managers, CIOs, and their various bosses make decisions and how those decisions are not necessarily rational. What you have here is an 18-minute seminar about IT decision-making featuring one of the world's most experienced IT industry analysts, who also writes a blog, Virtually Speaking, for ZDnet.

Dan Kusnetzky's Six "Customer Profile" Questions:

1. Please introduce yourself and your company
2. What were you doing that required this type of technology?
3. What products did you consider before making a selection?
4. Why did you select that product?
5. What tangible benefits have you gotten through the use of this product?
6. What advice would you offer others facing similar circumstances?

Robin Miller: I am Robin Miller, some know me as “Roblimo.” We are on the horn today with Dan Kusnetzky, a long time analyst and general interpreter of the IT scene, and where technology is going. So Dan, today, clouds – is it really cloud all the time, or is a lot of the cloud thing kind of fog-what’s up?

Dan Kusnetzky: If we go by what the National Institute of Standards describes as cloud, we need to include the thought that it is very similar to outsourcing workloads and then accessing them over the network from the remote data center. And the billing for it is as you use it, the management of it is self-service. And there are a number of categories of cloud computing. I have seen that suppliers have a tendency to want to ride on whatever the big thing is - whether it is big data, cloud computing, client server the list goes on and on.

And what they try to do is, whenever there is something that is remotely related to what they are doing, they try and attach that word to what they are doing. So if it is accessible over the world wide web, if you use a web browser as the interface - to some people that is cloud. I think a more strict definition would be that you would be taking a workload and moving it from your data center to operate in somebody else’s data center – you set it up self-service, you manage it yourself, you pay for it as you use it and that is a little bit more strict definition of cloud.

I have been seeing people offering applications such as SugarCRM, SourceForge, where they have outsourced a whole application – that seems pretty cloud like to me. There are other people who are offering machines upon which you can move your virtualized workloads, that is, you put your workloads into a virtual machine and you copy the file to their data center and it runs there. Those are everything from Rackspace to Amazon to a long long list. That seems to seem an awful lot like cloud.

But a lot of other people have been talking about cloud infrastructure when it seems for the most part, it gets slapped a web front end on what they are doing, so that people can go to their management and say, “I am buying a cloud”- I am reminded of a Dilbert comic strip, and this is ancient history back when relational database was the big exciting thing to be talking about, and the pointy haired boss came into the boss and said to Dilbert, “I want you to implement a database” and Gilbert responded, “Do you want it to be a red database or a blue database?” And the pointy haired boss said, “Blue database”. And so okay, we will implement a blue database.

What was clear was the pointy haired boss had probably read some article in a management newspaper or Forbes or some magazine and decided that they needed to implement a database. He had no idea of what they were, had no idea how they would be used, had no idea what it would cost, what it would do to help the company, but database was the key word to be able to say to his management, “Yes, we are implementing a database.” And I think cloud is kind of in that realm right now for a lot of companies - but not all. There are a number of people who are actually saying and I kind of classify this as a portion of the outsourcing trend where people have outsourced development, they have outsourced management, they have outsourced in some cases, they have put their own machines in somebody else’s data center so they can manage them and take care of them.

And this is the next step beyond that, where you are using somebody else’s machines, running somebody else’s licensed software, and you put your applications there and pay for it as you use it, kind of like the former notion of utility computing where you pay for it like you run a utility.

Robin: Okay, but I am going to give you a story about that. As a journalist, you know, we are outnumbered by PR people in IT by 20:1 or some such, so each one of us gets about a billion press releases a day. So I kept getting these utility computing database, utility computing databases, and I kept emailing back – “May I speak to one of your clients?”, “I want to talk to a utility computing client” - IBM could never dig one up, Oracle could never dig one up, Microsoft could never. They had no sales. Nobody was buying. But they were selling. And now seeing cloud, same stuff, same magical pixie dust - they are buying it. Why? Why is cloud better?

Dan: Well I think, one the infrastructure has improved. So it is possible to take a workload, encapsulate it in or more virtual machines, run it on a machine somewhere else on the network, have enough network performance where they would actually perform reasonably well, have enough storage locally to the systems running the application or components of the application that it actually can be made to work. And also the pricing is such that if I find an Amazon and I am buying a gazillion boxes, (gazillion is defined as some large but unspecified number – analysts have to define their terms!) they are probably going to be able to buy servers, software, power, communications at a much lower bulk rate than most small to medium sized companies who are going to be able to afford.

Robin: Yeah.

Dan: And then if you look at how powerful machines have become versus how much the workloads people put on them consume it makes a lot of sense to set up a multitenant environment kind of like a hotel

Robin: Wait a minute, we used to call that shared web hosting.

Dan: That’s part of it. A lot of this is a new name for what has existed, but the only thing that is really new is the ____6:59 that you use that the self service provisioning and paying for it by the use rather than on an annual or semiannual contract. But what we are seeing is, and I have spoken to companies who are implementing this for certain types of tasks either to reduce costs or to improve the time from the time they conceive of an idea to the time it is actually running, or to mitigate risk.

And mitigating risk is an interesting category that I see some people talking about, where people who have a great idea, but before they go to management and ask for a large amount of money to buy machines, buy software, find more space in the data center for their machines to run, get management people, security people, facilities, all of the things necessary to implement something – let’s just buy access to somebody else’s data center, try the idea out, and once it is producing something of value, then we can go to management and say, “We’ve been trying this out – it has only cost us this much to put this pilot project up; do you want us to put this into our data center with these limitations, you know, we got to pay this much for power and this much for floor space Or do we continue to just leave it at that supplier?” And it is one of those things where, if you take cloud computing, some managers will say, “Oh, that is really exciting, I will be able to go to my club, my golf club, my basketball club, and then say, “Yes, we implemented cloud computing, and it is saving us this much money.”

Robin: As a long term analyst who knows a little bit, a whole bunch about IT, are you telling me that people are making decisions on an ego base often instead of a rational business base?

Dan: You could say first just to say some background – I worked at a computer company for 16 years in various - software support, software consulting, product management, product marketing,business unit management; I’ve also worked at multiple research firms doing research on the market or managing teams of people that do research – if we look at the survey data and then compare it to actual behavior, if you ask people what they are doing, they almost always have a rational statement for why they did what they did. And if you look at what they said in the survey versus what they actually did, almost always they made a selection that was emotionally based. And then after the fact, came up with a nice rational basis for why they did what they did.

Robin: Heinlein quote time: “Man is not a rational animal; he is a rationalizing animal.” Now Robert Heinlein stole that from Mark Twain.

Dan: I was going to say ‘thank you Rob for Heinlein and Mark Twain’, yes. But I’ve seen that behavior over and over again because you ask people what type of automobile they are going to buy and they will give you a description of some reasonable fuel efficient vehicle that is going to last a long time, but is probably pretty boring. And then you see that same individual go out and they’ve bought a sports car that is red, and just consumes fuel and is not very practical, because it only has two seats, and it has storage about the size of a brief case - if you are lucky. Why did they say they were going to buy one thing, and do something else, is because you often find in surveys there are people who answer based upon what they think you want them to say rather than what they are really going to do.

Robin: Oh yeah.

Dan: Because it is really helpful when you are doing surveys to have really expert people construct the survey instrument, construct the interviewing method, so you can kind of get behind what they think they should be saying to you, and get to the data that you are really trying to get to – what are they really thinking. And quite often that means asking a whole bunch of questions that seem only vaguely related to one another; and what you are doing is getting puzzle pieces, and then the analyst at the end puts the pieces together and says, “They said these in the three comments but if you look at what they were going to do and their budget and their planning, this is what is more likely to happen.”

And so one of those things that analysts firms often have to interpret even survey results, which seem pretty cut and dried, before they publish the analysis and the final findings. And that is why, this is a pet peeve of mine, when I, like you, get email after email PR person trying to convince me to talk to a supplier. They often have tried the ploy of having done a survey of some type; quite often, the survey is a tiny sample, quite often, the sample was from people who went to one of their own events, or some industry event that was sort of vaguely related. They would sample, maybe talk to a couple of hundred or perhaps a thousand people, do an analysis, and then present this obviously self-serving statement that ‘the world is planning to do this’, when, at very best, you could say, the people who went to that event would that.

Robin: Well, there’s another factor to it. You and I, we are chronically married guys, we’re married, we just are, but there’s another thing that some of the younger guys I know use, well, they don’t tell you, but it is the CDI effect, have you ever heard of that?

Dan: I don’t believe so.

Robin: ‘Chicks dig it!’ That’s blinking lights, it is out there with those ooh lights shiny, and you got shinied into, you know what I mean, it is irrational but it is cute, it is cute. It’s the sports car.

Dan: One of the things that I have been doing as part of my research is a customer profile thesis. So I have actually developed a six-question customer profile which usually is short enough that I can get any executive to spend 10 minutes with me answering those six questions. And I have been publishing pieces on my own website kusnetzky.net, and some of them have also been cross published on ZDnet.com and my ‘Virtually Speaking’ column, but a lot of them are end users who have done something with cloud computing.

Usually the people I talk to hint at they were going for finding a way to live within their budget even though it has been cut back so severely that they can hardly operate, and by putting the computing in somebody else’s data center they don’t have the fixed costs of data center security for that data center, facilities, management, electricity, power, communications – they are sharing the costs with everybody else who is in that data center. So they don’t have to pay the whole thing.

Or they are saying, our own internal decision making process was so difficult to get through, and my need to get this accomplished was so great, that I went ahead and went out of the company to get something done. And most companies have a policy that more or less says, ‘it is easier to apologize for doing the wrong thing than to get permission to do it in the first place’. So they go on and they do what they need to do, and then present the successful results to somebody who pats them on the head and says, “You did a good job.” Only later they tell them how they did it.

And they may not even tell them how they did it unless someone has gone off of a budget somehow, but if you are under budget or within your budget, and you are getting things done, quite often people don’t question how you did it. So there are people who are movers or shakers, they’ve got work that needs to be accomplished, their infrastructure won’t support it, and they just go do something to make it work.

And that may mean going to Rackspace (I am just picking names, I don’t have any particular favorites) or maybe I will go to Hewlett Packard or IBM or somebody who is willing to offer access to somebody else’s machines in somebody else’s data center, and they can spin up a virtual machine, they can spin up storage, they can do what they need to do, and pay for only the part of it that they need. I’ve also interviewed some people using the same six-question thing that with spinning up a new remote office, and they couldn’t get new PCs and new PC images set up with all of the software that was required by their company; and so they basically went over the network to access virtual PCs.

And there are a number of companies who are offering desktop as a service. And they plan to do that temporarily and in some cases, they liked it and continued to expand upon this idea; in other cases, they didn’t seem to like it because of one reason or another and they pulled all of it back into their own environment. I have seen people start to use the cloud computing idea, and quite often, it is being adopted in the reverse order from what I’ve seen other technologies adopted over the last 30 years or so. Because usually, the previous pattern was research institutions, scientists would start with something when it was still even beyond leading edge - it was bleeding edge; it took a lot of work, it was very painful to use the technology, but their need was so great, and their resources to deal with that was so great, they could go ahead and use it when it was still really raw and make it do something for them.

Then it went to the people who supplied those people with technology. Then it went to large companies that hired people from those research institutions that brought the technology along with them. Then it went to their suppliers which were often medium sized companies. And small companies and governmental institutions were the very last to pick things up because to get through that whole chain on down.

Robin: But with cloud?

Dan: With cloud it seems to be in reverse order. Governments and small businesses who have been so severely their budgets have been slashed by the economy, and regulations and all kinds of things, and they need to get the work done or they are going to die.