Diving into cloud computing may not be as big a leap for your company as you might think. In fact, an important shortcut to such a move could already be sitting there in your data centers, waiting on standby to make the process easier.
"Virtualization is a stepladder to the cloud," says Laura DiDio, principal analyst with Information Technology Intelligence Consulting. So if you already have virtual servers running efficiently in your data centers today, then the road to seeing what cloud computing can do for you will be far smoother, she says.
"Virtualization is the first step," DiDio says. "It gets you in a cloud state of mind and it gets you thinking about things that are important, like easier maintenance, lower costs, improved reliability and better scalability. In terms of using virtualization, companies are further along with that than they are in using the cloud. Virtualization is more mainstream at this point."
That progress happened as businesses moved toward massive server consolidation projects in the last 10 years or so to cut costs, reduce maintenance and lessen data center floor space needs. Virtualization might also have been a bit scary at first glance, but it has taken off and provided lots of benefits for IT departments and corporate bottom lines.
"So cloud computing is a natural outgrowth of virtualization," DiDio says. "The idea is that you're sharing things. It's using multiple servers on a digital network as though it were one computer. When done correctly, except in very rare instances, all cloud environments are going to feature virtualization, but not all virtualized environments will be cloud based. Some companies will still only run their applications in their own data centers."
Since 2000, virtualization has been evolving, she says. "Those first virtualization 1.0 implementations were fairly straight-forward and they delivered immediate results, particularly at that time after the dot com bust. Many users said they were able to cut their hardware costs in half. As time has gone on, we've seen virtualization 2.0 and 3.0 and we're now seeing organizations move up the virtual stack. Instead of just virtualized applications, they're also going to desktop and storage virtualization."
With all of that in mind, DiDio has five key tips for making your company's transition to the cloud less worrisome and a whole lot more successful:
1. Before deciding your strategy, ask what benefits you hope to secure by moving applications and processes to the cloud.
"You have to help frame the discussion about what is being sought," she says. "Those are typically reducing costs and management headaches, easier provision of applications, better reliability and improved scalability. You need to determine the specific functions and features that you need to support your critical apps."
2. Identify and discuss the major challenges of the cloud for your particular business.
"Obviously security is huge for certain verticals, as well as compliance in industries like banking and healthcare, DiDio says. "You have to look at what you do and what you need."
3. Find the business model that's right for your users and your business.
"You have to decide whether you will use public or private clouds or a mix of the two, then you have to decide on your cloud delivery model," she says. "Are you going to do SaaS, Platform-as-a-Service or Internet-as-a-Service? Most go for SaaS for the applications that they run on the cloud. Most companies choose to have someone else do it for them."
4. Find out where the data is and be comfortable with how it is handled by your cloud vendors.
"The issues you have to be concerned about are the viability of your vendor, their tech support and their response time," DiDio says. "You also want to ask about latency guarantees. You don't want anything to lag.
"The biggest concern that people have when they go to the cloud is about where their business data actually resides," she says. "If you can't see the data yourself, you have to be very, very confident in your cloud provider if you've got a public cloud." Be sure to ask what the vendor's response time will be in the event of service problems. Also ask about how the vendors will meet your requirements for up-time and service-level agreements.
You also need to know about the other tenants who are sharing your servers in the vendor's cloud, she says, so that your applications and performance aren't affected by the work they are running.
"In a public cloud, you are talking about a multi-tenant environment. You need to worry about your neighbors and you need to find out from your vendor who else is in there with you," she says. "You want to ask if you are going to get your fair share of the resources or are you going to be in there with someone who will hog your resources? You want to find out as much as you can."
5. If you want to maintain a private cloud for security and control, then you have to invest in training and staff.
"We talk with a lot of clients who are doing private clouds," DiDio says. "More are going with private clouds than public clouds. That sounds great, but to make it really work well, you are going to have to invest in training and certifications ."
The problem is that many companies reduced headcounts on IT staffs during the economic recession of 2007 to 2009, and many are still doing more with fewer people and smaller budgets. "One of the first things that went out the window at that time was money for training and certifications. Now with a cloud initiative you're tasking your IT department to provide higher levels of uptime and reliability, easier and faster provisioning, as well as better security. If you're essentially installing something like virtualization and the cloud to meet these goals, you have to make sure that you get the necessary training for your IT staff and that you review your existing infrastructure.
In addition, you will have to figure in the need and costs of newer hardware. "If you're intending to put newer applications on five year old servers, that could be problematic," she says. "You want to make sure you have up-to-date hardware and software, enough to carry the planned workloads. You have to make sure that you correctly configure the virtual and cloud environments."
You also need to be sure that you plan and build disaster recovery and back-up provisions into your cloud strategy, she says.
Once you've decided on these first steps, pause before you begin implementation.
"Don't rush the deployment," DiDio advises. "This is where you can get a disconnect & from the users who will be working with the applications on the cloud. You cannot ignore the human issues. That means that you have to have the CEO, CIO, CTO, COO and all the appropriate people in IT, telecom and even facilities together, sitting down and discuss these issues."
This is the time when you need to listen to your IT managers about the experiences they've been having with vendors, DiDio says. You might hear that though a vendor is offering the best price they are not delivering in the area of support or giving sufficient attention to the business needs due to a recent acquisition.
It's the time you spend here, before any vendor contracts are ever written or signed, that will really pay off in the end.
"You have to do as much research as you can," she says. "These technologies of virtualization and the cloud offer tremendous potential, but you don't just wave a magic wand and all of this magically happens. You've got to take your time and do it right, and like any other technology, it will have its upsides and its downsides."
This idea of virtualization as a perfect lead-in to cloud computing is a recent but growing trend, according to DiDio. Major IT vendors are leading the way in this direction, from Dell to Hewlett-Packard to IBM, Microsoft and others, she says. "They are intertwining their virtual environments so they are stepladders to the cloud. Microsoft now has versions of cloud-based applications such as Azure and SQL Server. Both sides, users and vendors, are heading to the same place."
Sign up for Computerworld eNewsletters.