Skip to content

A Little Cloud Heresy

by on February 10, 2016

“We are at the beginning of the age of planetary computing. Billions of people will be wirelessly interconnected, and the only way to achieve that kind of massive scale usage is by massive scale, brutally efficient cloud-based infrastructure.” – Dan Farber, Editor in Chief CNET News

What if the cloud really doesn’t change everything about computing from this point forward?

Business managers love the idea of the cloud. IT has always been a bit of a pain. Wouldn’t it be great if they could spend less and get everything they need without any hassles? These are the same business managers who love outsourcing in general. Every few years some big outsourcer convinces them that they can save money on IT by outsourcing to an IBM, Dell, or HP. 3 years later, they’re desperately trying to get out of the contract as they realize in the end they’ve spent much more and gotten far less.

Hardware manufacturers are challenged by the whole cloud thing. Dell and HP have given up on providing their own public cloud options. Turns out that buying a server to put in the middle of a datacenter somewhere in Texas just isn’t that compelling in comparison to having it in your own data center, unless you just have a generic jones on for an outsourced data center. Depending on your needs, moving to a hosted data center isn’t a bad plan. Gotham uses one. But honestly, not world changing.

Software companies are generally embracing the cloud thing. Why wouldn’t they? To make software work you need to buy compute and storage. Depending on the nature of the software, you might typically spend 2-3 times more for the hardware to run the software than you did on the software itself. Why should SAP, Microsoft, or any other software developer put dollars in the pockets of Dell, HP and EMC? Cloud software keeps all that revenue with the software company.

While this can be a good idea, it isn’t necessarily always utopia. For one thing, it doesn’t seem like cloud software is fundamentally cheaper than an in house deployment. Microsoft has admitted that Azure isn’t really fundamentally sound economically. If customers actually used what Microsoft sold them, they’d be losing money. They’ve got enough people on so that I wouldn’t really expect some sort of breakthrough coming from economies of scale at this point. It seems like the plan is get everybody to the cloud and then regain margin by raising prices. Just saying.

Economics aside, the technical perspective on software as a service is pretty simple. In some cases cloud delivery works great and there’s not of a lot of integration with other systems. These are great cloud solutions. In other cases, cloud delivery just doesn’t work well or there’s a need to integrate the system with a large number of other systems. In these cases, cloud just isn’t that great an option.

Until we defeat the laws of physics once and for all, the physical location of our data and our applications is going to be important to us. It’s important for performance reasons. It’s important when we’re integrating systems. And it’s important when we’re doing cross platform data analysis across our organization. And, while we’re talking about laws, we may as well talk about legal issues as well. Data sovereignty issues are not going to away. You may not care where your data is, but every country seems to care a great deal about where its citizen’s data resides.

I know a lot of people see security as the Achilles’ heel of cloud but I don’t really think it’s the biggest issue. Encryption technologies can help a great deal. I honestly think that there are secure enough ways to do cloud for nearly every use; they just might be more expensive than you would want them to be.

Which leaves us with the golden child of the cloud, Amazon. Amazon offers a great service and seems to be doing it at a profit. Have they found the secret? In some ways, they certainly have. First of all, Amazon has a very low cost delivery model. They’ve built some excellent systems based on open source software and commodity hardware. Secondly, they’re not cheap. They’re prepared to charge a fair amount and if you don’t like it, you can move on. Many startups start at Amazon and move on to their own data centers as they get more predictability in their computing needs. That’s fine for them, and it’s fine for Amazon too, as there always seem to be new takers for their cloud. For me, the point is the same – it’s great for some people. If your workload needs significant elasticity above all else, Amazon is a great place for you to live.

It was pretty common knowledge in the early 2000’s that IT workers wouldn’t exist in this country by now. Globalization was killing the US IT industry and soon all technology work would be done from India. Globalization is a great outsourcing option for many types of work. Did it kill all other models and reign supreme over all of IT? No. In fact, as the word outsourcing came to signify off-shore outsourcing, a number of hybrid solutions evolved – right sourcing, near shore, etc.

Look, you can open all the great restaurants in my town that you want, I’m still going to want a kitchen in my house. Cloud technologies are going to be a great solution for some problems. The inevitable, complete, perfect, and final solution for IT? No.

Comments are closed.

%d bloggers like this: