How To Transition To a Cloud-Based Infrastructure
Poll CTOs about game-changing technologies they’re watching, and I’ll bet the vast majority will put cloud computing at the top of the list.
While cloud computing isn’t yet mature enough to put your entire IT infrastructure on, there is an undeniable paradigm shift in what constitutes a company’s “data center”.
Below are some frequently asked questions and answers that will get you up-to-speed on the cloud computing revolution.
I. What is the “cloud”?
NetworkWorld defines cloud computing as a combination of several existing technologies, including:
a. Grid computing – multiple commodity servers are tied together to form one big virtual server. If any of the individual nodes fail, the grid remains unaffected and the user is not affected.
b. Utility computing – metered computing, where you are charged based on usage. Pay for your computing like you pay for water or electricity.
c. On-demand computing – most cloud services can be managed from a web interface, and users can provision more or less computing power with the click of a mouse. This feature is especially useful when you need to temporarily scale up for a big marketing campaign.
When people talk about cloud computing, they are generally talking about renting cloud services from 3rd party vendors (ie., public clouds). However, you can also build your own private cloud computing infrastructure in-house to replace separate, individual servers.
II. So what? What are the business benefits of cloud computing?
According to the Open Cloud Manifesto:
The key characteristics of the cloud are the ability to scale and provision computing power dynamically in a cost efficient way and the ability of the consumer (end user, organization or IT staff) to make the most of that power without having to manage the underlying complexity of the technology.
This shift from managing individual hardware nodes to managing a cloud means CTOs can spend more time (and money) on strategic, rather than tactical, work.
As Lew Moorman of Rackspace’s Mosso has said:
Currently businesses spend 75% of their IT budget on maintenance, 25% on strategic activities. The real promise of cloud computing is inverting that equation.
Benefits of offloading your IT infrastructure to the cloud has several benefits:
a. Reduce costs. By taking advantage of the scale of 3rd party cloud providers, you’re sharing the burden and cost of hiring, training, and managing hardware and OS gurus with other customers.
b. Let someone else deal with the underlying technical details. Instead of comparing hardware specs and prices for your next round of server purchases, or skimming hundreds of resumes and conducting hours of interviews to hire a crack team of sysadmins, you can worry on big picture issues. Like where to apply your IT dollars where they can do the most good for your company’s bottom line.
c. Be leaner and more agile. When you don’t have hundreds of your own servers to worry about (and the big team of sysadmins required to keep them running), you also don’t have worry about hardware sitting idle or having to layoff IT employees when business slows down.
III. What can I use the cloud for?
Storage and CPU cycles. In theory, anything you would run on servers in your own machine room can be run on the cloud.
However, in practice, there are limitations to what you can export to the cloud. For example, legal regulations may force you to keep sensitive customer information within your firewall, or licensing or technological limitations may prevent your company from using cloud services.
IV. How much does it cost?
Vendors typically charge by metering usage.
For example, Amazon charges between $0.10 – $0.80 per Linux server instance hour or $0.12 – $1.20 per hour for a Windows server on its Elastic Cloud Computing (EC2) platform. On the storage side, Amazon’s Simple Storage System (S3) charges for the space ($0.12 – $0.15 per Gb per month) plus transfers ($0.10 – $0.17 per Gb) plus requests ($0.01 for 10,000 GET calls).
V. What are some problems with a cloud computing strategy?
The cloud computing industry is still in its infancy, which means most of the rules, standards, and best practices still being hammered out.
Here are four problems CTOs face when exporting their infrastructure to the cloud:
a. Security. Cloud services typically have multiple customers sharing the same physical hardware. (It’s like being on a giant shared hosting machine.) While virtualization mitigates much of the risk of shared hardware, nevertheless, you are sharing server resources with other customers and may be at the mercy of other people’s crappy code. With an in-house data center, at least you know where your own vulnerabilities and dangerous hacks are.
b. Interoperability. This is currently one of the biggest drawbacks to using clouds. The industry needs to work together to solve this problem if they expect CTOs to flock to cloud services en masse.
As Robert Grossman, managing partner of Open Data Group, said in March 2009, “it was clear… that there was no consensus yet about how to begin to draft standards for clouds and for their interoperability. This was not surprising. But what was surprising, at least to me, was that there was not even consensus about what were the real problems and challenges.”
So there is still quite a ways to go before industry standards are established. By the way, this situation is an opportunity for CTOs to help shape the cloud computing standards that will inevitably be established in the next few years. Check out the Open Cloud Manifesto and the Open Cloud Consortium for the current state of cloud standards discussions.
c. Software licensing. Traditional software licenses do not fit the cloud model. When your hundred-server render farm suddenly becomes one giant virtual server, does that mean you can reduce your software licenses from 100 copies to 1 copy? Probably not. Software vendors are scrambling to develop new licensing models that can work within the cloud paradigm, but they’re not there yet.
d. SLA details. What does “100% uptime” really mean? Some vendors promise “100% uptime”, though a careful reading of their SLA (which requires having your attorney interpret confusing legalese) may show that your idea of 100% may not be the same as the vendor’s idea of 100%. Clearly explaining what they will do when the inevitable downtime occurs will be critical if cloud vendors want to earn CTOs’ trust.
VI. What’s coming up on the horizon?
a. More vendors. Amazon Web Services (AWS) is currently the largest public cloud vendor in the market, but lots of technology vendors are developing their own cloud products. Major hosting companies like Rackspace (under the Mosso brand) and Joyent already have some kind of cloud offering, and other technology players are joining the game. Look for cloud-based products coming soon from IBM, Microsoft, and Sun.
b. Better technology. Customers are telling IBM that “we want a shared infrastructure, but we don’t want it as open as Amazon,” says Kristof Kloeckner, IBM’s cloud computing CTO. Look for features that protect your cloud infrastructure from other people’s mistakes. Also look for better safeguards against the kind of irrecoverable data loss that was widely written about by the tech press last year.
c. Industry standards. Standardization of protocols and tools that work with multiple cloud vendors is a requirement before cloud computing becomes ubiquitous in the enterprise. Getting your data on and off the cloud is, unfortunately, currently dependent on vendor-specific protocols. And cloud products from different vendors do not play nicely with each other. For example, it’s still way too hard to, say, transfer data from one cloud storage system to another one from a different vendor.
V. Where can I learn more?
Read the Open Cloud Manifesto to understand the existing problems with enterprise cloud solutions.
Subscribe to InformationWeek’s Plug Into the Cloud blog to stay up-to-date on news and developments.
Subscribe to Jason Baker’s Cloud CTO blog. You’ll find links to news and commentary on cloud computing developments and the industry from the CTO of a cloud computing company.
Join the Open Cloud Consortium as a member, participate in the working group discussions, and have a say in the development of cloud computing standards.
