A Cloudcamp was held in Toronto on April 6, 2010. For those of you who are not familiar with the concept of a camp, it is usually an unconference: there is little or no traditional presentations (one active presenter vs. a more or less passive audience). Instead of that, most camps are unconferences where people just come face-to-face and share their knowledge. These notes are a somewhat shortened record of what was being said that day. (See the links for the sound files in the end of this article.)

In Toronto, we have camps for everything. There is a BarCamp, a BookCamp, a PodCamp, a ProductCamp and even a CupcakeCamp (I am not kidding! Definitely worth a visit if you have a sweet tooth). So, the CloudCamp was about cloud computing. (Wikipedia defines cloud computing as “Internet-based computing, whereby shared resources, software and information are provided to computers and other devices on-demand, like a public utility”.)

For the unpanel, the facilitator (Dave Nielsen) asked the crowd who they thought were cloud experts and then invited these experts to sit in front of everybody as the panelists.  People came up with the list of questions and the panelists got to choose a question and had one minute to answer it.

Now, most people just aren’t very far along the cloud deployment route. People are still trying to figure out things like security, lifecycle management, ROI, types of cloud, vendor lock-in, future of cloud, etc. The camp attendees varied from seasoned “clouders” like Reuven Cohen of Enomaly to complete newbies. Those are the questions that were offered by the audience:

  1. How do I do App Monitoring?
  2. How should I change how I design my app?
  3. How does Cloud affect the performance of my app?
  4. How do we bring Cloud to the mainstream?
  5. How do I address management concerns about security?
  6. How do I do automated provisioning?
  7. What are the key ROI measurements?
  8. How is software licensing different and is it possible to be non-compliant
  9. Are there standards for cloud?
  10. How do I start to take advantage?
  11. What is the future of Cloud Computing

Not all of them got answered but some certainly did, and the answers given were very interesting.

(Note: I did not catch the names of all the people who answered questions, so if one of them is you, please let me know so I can acknowledge it properly! – Tania)

Question #5. How do I address management concerns about Security?

- My name is Omkhar Arasaratnam and I’m a Sr. Security Architect with IBM. I lead the “Security in the Cloud” project for the Open Group, so I have a bit of experience with this. The main issue around security in the cloud currently, in my opinion, is there is a lot of uncertainty. Inherently there is a lot of unsecurity because we don’t know what we are getting into. Where customers start to get a little more concerned is about the governance behind it. So if I have a developer who has gone rogue, do you have the governance in place that says we have advocated that you can do this sort of stuff vs. that sort of stuff. How are you embracing it. How do I incorporate that new stuff into my corporation in a safety net. So the security concern while a valid one, is more of an issue of how you adopt technology internally, rather than it is cloud and it is scary.

Question #7 – What are the key ROI Measurements?

I’m Tim Smith, I’m a co-founder, former systems hacker and CEO of GridCentric which is a local platform company here in Toronto. I picked #7 because I think everybody gets that it’s a transference from capital expenditures to operating expenditures. So the Azure guy showed about the conversion from upfront capital expenditures. Up front capital expenditures, which means you buy a bunch of servers before you actually get your customers in  and sometimes you buy too many or sometimes you buy too little. But the thing is that you have to buy them before they come. You know, you have to build it before they come. When you move to cloud infrastructure, you’re changing that to operating expenditures, which means you’ve got one customer, you have one server. If you have a hundred customers you have two servers, if you have a million customers you have 10 servers.

You’re going to scale your assumption based on how much actual demand there is for the service. So that’s been the kind of classical ROI change or the classical value driver, is the conversion from cap ex, from upfront expenditures, to a “pay as you go” model. One ROI factor that people don’t really think about a lot is the ability to actually… the value you get from being able to handle sudden bursts in traffic, which is really a reduction in opportunity cost. If you’re a web based start-up, for example, and you’re running on your own equipment and suddenly you get a demand burst, you get slashdotted or tweeted and all of a sudden you have a billion people coming to your website, you can’t run out to the store and buy new servers and plug them in in time to service all that demand. Whereas if you’re using cloud infrastructure, ostensibly you can increase your server count automatically and quickly and if your business model is structured right, the more customers you serve, the more money you’re making. So you are very easily able to capture that value.

Dave Nielsen:   So less money up front, less risk and being able to take advantage of the opportunity when it comes.

Question #2 – How should I change how I design my application?

There’s kind of two buckets that I think about these things. One is labeled IAAS, Infrastructure As A Service, and one is labeled PAAS, Platform As A Service. And which bucket you fall into and which one is right for your business is often a decision that takes into account what your team strengths are, what type of problem you’re trying to solve, how you plan to grow. The best way to describe this in a minute is for people that are used to thinking about boxes, I’m used to thinking about a server as something that does a job, I put it on a shelf and that’s now my database server. Or I put this one on a shelf and that’s now my web server. Those people best map onto the Infrastructure As A Service clouds because it maintains those same paradigms. And you don’t have to change a lot about how you architect your application if that’s how you’re already thinking in terms of the infrastructure at core. There’s different things that you do want to change in terms of how you scale it, how you automate it, how you pull back resources, how you have to manage multiple facilities. Because now that’s a lot more feasible than it used to be before. Before you had to get to an economy of scale to go from one location to two locations, to be able to support that. Now you can just buy little pieces of infrastructure all over the world or wherever your users are and it’s a lot more feasible to do that. On the other side, if you’re not used to thinking about things within boxes, being confined within  2 GB of RAM and 100 GB piece of disk, you’re more used to thinking about I want to deal with requests per second, I want to deal with how many things I’m going to process. Well, the Platform As A Service best fits that because you’re abstracted away from the boundaries of that box. So you kind of scale seamlessly without adding new pieces, even though it’s virtualized, even though it’s automated, it scales a little more seamlessly. But crossing over between those two tends to be a bit of a challenge and it tends to be a bit of a challenge because wherever those strengths initially developed, you now have to rethink how  you design your application.

Question #11 – What’s the future of the cloud?

So, how many read “The long tail”? So the long tail may be that there’s somebody in South America making weird baskets and maybe there are only 10 people in this whole village that even liked it. Well, now there’s 10 people in 1,000 villages across the world, right? And so the reason why the cloud is looking really good, because the long tail has exposed an unbelievable amount of consumerability that we’ve never been able to see before. And it even gets better because how many people have read Malcolm Gladwell’s “Tipping Point”, right? So you combine that, it’s not a long tail, it’s like a stegosaurus, right? And how many people heard of FarmVille? It’s probably the biggest data center on the planet right now. It’s the fastest growing data center in the history of mankind, right? It’s a Facebook application, right? Etsy, which is another company that didn’t exist about a year ago. They sell arts and crafts in an eBay style, right? So the stegosaurus is just creating the opportunity for just phenomenal growth in infrastructure. Because of the Cloud, the barrier to entry to owning a data center is gone and you’re going to see more and more pop up infrastructures that are just going to grow out of nowhere and it’s just going to consume the cloud because it’s the only way that they will start up. And anybody in the stegosaurus spike is going to start off based on what you were talking about, cap ex and op ex. So there.

Question #3 – So how does cloud affect the performance of my apps?

Basically the difference is you’ve got a bunch of machines that are either managed for you or you are facilitated in managing them yourself, which is sort of the Tom Sawyer Cloud Company, or you decide that cloud is bad, and you’re going to stay at your cubicle close to your computer where your high level of security is soon going to be hit by a tornado or an earthquake. So I think the geographical distribution of cloud is going to improve performance because we can swizzle out cloud instances at will if you’re willing to pay for them and sometimes even if not. You can get enormous performance and peak aligned, so 24-hour, as needed performance on an arbitrary number of instances. And whether you do it yourself or you get the help of a company, as mentioned earlier, RightScale who will keep track of how many servers you’ve got going, kick another one into production if needed for load. It’s pretty cool.

Question # 10. How can I start up and take advantage of the cloud?

Yeah, that’s the real question. All this other stuff, ROIs and stuff, all leads back to how am I going to make money with the cloud? So I think as a start up there, there obviously is certainly a bit of a gravy train happening, everybody’s talking about. Look at  the Cloudcamp, for example. We created it what, two years ago and now we’re what, in 70 countries or cities or whatever? It shows that there’s a demand for anything cloud related and the problem is most people have no idea what the hell cloud is, right? So it’s this all encompassing buzzword that means everything and nothing at the same time, which is sort of a double edged sword in a sense, that you can ride this sort of buzz, hype thing that exists out there to your own purposes from purely a marketing point of view.

So what cloud computing does for start ups is one thing; it allows anyone with a credit card in their basement to compete with Google or Microsoft or IBM because now I don’t need to create the world’s largest infrastructure in the world, I can use someone else’s who is probably better at it and has already done it for me. And all I need is a good idea. So the problem, again, is coming up with a good idea but the big opportunity is the fact that I don’t have to worry about the nuts and bolts of making my good idea a success. Whether that’s FarmVille or something else, Twitter, I can use someone else’s infrastructure… and do what I’m good at, whatever that happens to be. So again, the opportunity, I think, is to find those particular niche areas and make a buck in cloud computing using cloud computing infrastructure. I’m not sure if that answers it at all or if this is a random rant, but…

Dave Nielsen: No, that’s good. So yeah, the point is yeah, let somebody else do all that heavy lifting with the hardware. You can focus on your special area of expertise. Very good, okay.

The problem is the integration, right? But go ahead.

- So I was going to provide an all encompassing caveat that sort of goes for everything that we’ve said here. Which is, cloud’s right for you provided you want to take that business risk. And this gets back to making your dollar, right? So you subscribe to… Your integrator subscribes you to, we’ll call him a cloud general contractor, right? He’ll get you your infrastructure, whatever it is, whatever it is you need to do your business. And that cloud goes dead, what’s your recourse? A refund on your credit card from cloud provider?

- There’s also additional exposure. If you’re on IAAS, multi-tenancy. Our friends to the south have these organizations called the FBI, etc., who kind of kick down the door, and have enough warrants, they pull out a server and… You know, the FBI isn’t graceful about it when they pull out boxes, right? They don’t give a damn whether you’re the only tenant on there or if there are several other tenants on there. Forensically they want the entire thing. What happens then? What happens if your cloud provider opens up something that’s in hostile territory in North Korea and you’ve got data residency issues? So there’s a lot of things for a business…

- I know there was a hosting provider in Texas which wasn’t a cloud and they shut down the whole hosting project. So that’s not a cloud problem. They shut down the whole hosting business and everybody got locked out of their business for six months. It’s not the cloud.

- Right, agreed. So what I’m saying is whenever you make an IT sourcing decision, be it traditional “buy your own box”, be it “throw it in the cloud”, be it an integrator hybrid, you or whoever makes the business decision in your company about profitability needs to be aware of the risk. With anything there’s a risk. Like my colleagues have said, you host it in your basement, power goes out, you’re screwed. So what is your risk, what’s the reward, what’s your mitigating control? That’s what it comes down to for any of these things.

Question # 9. Are there standards for the cloud?

So for application portability, I’m running my application on one cloud, I want to run it on another, okay? The higher up in the stack you go, Infrastructure to Service, Platform as a Service, Software as a Service, the higher you go up the more difficult it is to import, okay? But you know, there’s really… The standards at the top level are usually around data because you’re not going to take somebody’s Software as a Service product and run it in your own infrastructure. That doesn’t happen. If you export the data it’s totally different. Maybe you can export the data at the bottom layer. It’s not so much about data, it’s more about the applications running on a virtual machine, can you run them somewhere else, right? So it’s kind of different at the different layers. So are you talking more about the infrastructure? No, you were talking about the Platform as a Service layer.

- Yeah, I’m thinking largely of that, that you have to go to the infrastructure layer if you really want to get any portability?

- So the lower down you go…

- Well, you’re going to be locked in anyways and you’ve just got to choose where you’re going to be locked in. But in Google App Engine for example, you can program in Java, there’s a bunch of frameworks available for it. So for instance, if you use Django as your framework then you can move it. The problem will be yeah, database related because you then have to figure out a way to use their database structure outside of their environment. And I think there are several open source projects that are geared to do just that. So I think if there’s a strong enough demand for any application, you’ll find an ecosystem emerge around it to provide the tools that allow you to move from it. So whether it’s Amazon from an infrastructure point of view or Azure or Google App Engine or SalesForce, there are a variety of tools available. The problem is if you choose a platform which isn’t broadly adopted, then you’re going to have a lot… I guess much less options.

***************************************************************************************

Here are the slideshares of some presentations:

Here are some Audio Recordings by Session (User ID: audio; Password: cloudcamp):

2nd session: CloudCamp 2. Cloud Computing: Return on Investment