Archive

How I convinced my imaginary friend that cloud is cheaper

Share this post:

I never had an imaginary friend as a child—apparently imaginary friends require having an imagination, which more or less disqualified me (I’ve always been able to quickly add two and two together, but if you give me a blank sheet of paper to create something, I’m lost!) However, I created an imaginary friend just for this blog post. Here are some of the silly things he said about IT projects and cloud computing, along with my responses.

I tried to do this as a kid but it never worked.

1)  “We need the largest server available for my project!”

I don’t think it will surprise anyone that many people overestimate the capacity requirements of their projects. One reason for this, as with many things, is self-interest. People are more likely to get into trouble if they underestimate load and the application performs poorly (or crashes!) than if they overestimate and spend too much money. In addition, I think there is a psychological element of wanting to be part of something big and important. (You may not want to take psychological advice from a crazy IT architect though.)

Cloud computing helps combat this overestimation (and legitimate worst-case planning) because infrastructure can be quickly scaled up to meet demand—by humans and perhaps also by algorithmic agents when certain conditions are met or certain forecast data is available. (The algorithms have already taken over the stock market in the United States, so it’s not a big stretch to see that they will also begin making decisions in other areas.) So, quit shaking your Magic 8 Ball to generate load forecasts. Design the application to dynamically make use of cloud infrastructure so that it can accommodate a much wider range of loads.

2)  “My server will be steady-state. Is provisioning it on the cloud really cheaper than bringing it into service manually if it’s going to stick around for several years?”

It’s true that cloud (particularly infrastructure as a service) can help more in dynamic environments, although as stated above I think even long-term production environments should be more dynamic. However, there are other concerns in play besides just the initial provisioning. If everything is provisioned through the cloud, then service management processes, reporting and so on can all be fed automatically by data from the cloud systems. If you’re still manually setting up infrastructure, you’re also manually pushing that information into those processes and reporting systems, with the possibility of human error.

Besides, depending on the cloud solutions you choose, in many cases it won’t cost any more to provision new workloads on cloud (although upgrades and migrations of existing workloads might be a different story!). For infrastructure as a service, after virtual machines are provisioned, in many cases they’re basically just like any other machines that you created the hard way. Of course, the details depend upon how a private or public cloud is configured, but there are a lot of options. If there’s a need, you should be able to match the characteristics and service level agreements (SLAs) of your traditional environments. For some good advice on negotiating SLAs, you may want to look at the Cloud Standards Customer Council’s “Practical Guide to Cloud Service Level Agreements.”

3) “It’s cheaper to buy than to rent!”

Okay, my imaginary friend may have a point with this statement, but this is the exception that proves the rule. Also, the correct end to that sentence is, “… if you’re using it all of the time.” Think of it this way: a large company may have a group of lawyers who are employed full-time. (And trust me, you do not want to annoy them. They’re even worse than the accountants!) However, this same company may hire outside law firms to take care of peak workloads and help the on-staff lawyers instead of hiring more permanent staff.

In computing, if you’re fully utilizing a certain amount of infrastructure all of the time and you have the skills in-house to run it, owning the infrastructure might be cheaper. In this case, however, hybrid cloud can still be the cheapest way to go—you provide your own “base load” as part of your private cloud and then spill over “peak load” to a public cloud provider. In addition, you could also implement business rules to set which workloads are allowed to go where. For example, if the application has the “super secret” rating, it always goes on the private cloud on dedicated hardware inside your firewall. Also note that many cloud infrastructure providers have “reserved” pricing, where you get a considerably better price for committing ahead of time to use the infrastructure.

Also, don’t forget about having “skills in-house to run it.” The hardware is cheap, but how much does it cost you to staff a large IT organization that knows the best practices for running a data center?

Is cloud cheaper?

In many cases, I think it will be, primarily because of workload consolidation and higher utilization. Many organizations don’t need to run a server farm or a huge data center any more than they need to have a lawyer or accountant on retainer—it’s a service that they “rent” as needed. And, like a shared cloud, many law and accounting firms are also “shared” among multiple clients and may have access to sensitive information. The cloud model has worked for those professions for many years as a secure and less expensive option to doing everything in-house.

Do you think cloud is cheaper or not? You can leave a comment below or reach me on Twitter @crdotson. My imaginary friend does not yet have a Twitter handle.

More stories

Why we added new map tools to Netcool

I had the opportunity to visit a number of telecommunications clients using IBM Netcool over the last year. We frequently discussed the benefits of have a geographically mapped view of topology. Not just because it was nice “eye candy” in the Network Operations Center (NOC), but because it gives an important geographically-based view of network […]

Continue reading

How to streamline continuous delivery through better auditing

IT managers, does this sound familiar? Just when everything is running smoothly, you encounter the release management process in place for upgrading business applications in the production environment. You get an error notification in one of the workflows running the release management process. It can be especially frustrating when the error is coming from the […]

Continue reading

Want to see the latest from WebSphere Liberty? Join our webcast

We just released the latest release of WebSphere Liberty, 16.0.0.4. It includes many new enhancements to its security, database management and overall performance. Interested in what’s new? Join our webcast on January 11, 2017. Why? Read on. I used to take time to reflect on the year behind me as the calendar year closed out, […]

Continue reading