Smart businesses don’t spend a penny before clearly understanding what they are setting out to accomplish. This includes when considering how cloud fits into the business, especially when big data is a part of the equation.
Deciding what you want to accomplish with cloud computing is just as critical as actually doing it. You need a cloud strategy.
So how do smart companies proceed?
Here are some tips for creating a successful cloud strategy.
Decide What You Want
Start with the end in mind by considering the following list of outcomes. What you select here determines where you want to go with cloud, and what its value is to your organization:
- Cost savings. Saving money is always high on the list, but with cloud you must choose your spots. With Infrastructure-as-a-Service (IaaS), for example, cost savings is a real possibility, as those prices are falling. For Software-as-a-Service (SaaS), on the other hand, this may not be the case, as SaaS prices tend to be rising.
- Agility. You want the ability to innovate and change at a faster pace than you can in your own shop.
- Speed. Related to agility, you need results quickly.
- Delegating responsibility. Maintaining all of your IT infrastructure 24/7 is a burden you’d like to shed.
You may think of a few others, but these four are a good place to start. Choosing outcomes brings clarity on what matters and provides a way to measure success.
Having selected the outcomes, your next task is to prioritize them, as that will help you determine where to start.
Select your Approaches
When deciding how to implement cloud in our business, there are four approaches to choose from:
- Cloud Never. Situations in which you would never consider the use of cloud services. For example, systems with regulatory-compliance issues that flat-out disqualify cloud.
- Cloud Maybe. Situations in which you might use the cloud if you can get a better outcome than doing things the non-cloud way.
- Cloud First. Situations in which you will always look to the cloud first before building it in-house.
- Cloud Only. Situations in which you will use only cloud and never build it in-house.
It’s important to note that you are not picking just one of these as a one-size-fits-all approach. An overall strategy will include any or all of them, as appropriate to your needs. For example, you might adopt a cloud-only approach for email because it’s become a very mature cloud offering. On the other hand, you might go cloud-first for office-productivity and collaboration applications. An example here would be Office365, which is a robust solution for the standard Microsoft Office apps but not so much for SharePoint. Big-data services might be cloud-maybe, depending on your existing in-house capabilities.
Easy Does It
The term “best practice” can always be defined as “avoiding the worst practice.” In the case of cloud strategy, that boils down to this: Cloud should be adopted one service at a time.
Trying to move too many things to cloud at one time is a recipe for disaster. It’s better to start with the low-hanging fruit, such as finding a file sync-and-share service (e.g., Dropbox, Box), or perhaps a provider for compute (IaaS). Backup and disaster recovery are also good starting places. Selecting a Hadoop or Spark service is yet another. Make decisions for each service individually, including how you will migrate to them, how you will get out if necessary, and how you will integrate them with your existing systems.
It’s worth pointing out here that cloud computing is not a data center. You are not moving your data center to the cloud provider, you are changing how you are acquiring certain IT services and functions. (That isn’t to say you should never move your data center to a colocation facility. The point is that such a move would be data-center outsourcing, not a move to cloud computing.)
It’s important to be careful of blanket all-in-one strategies, or statements that simply aren’t strategies at all, such as the following:
- My strategy is (pick a public cloud…Azure, AWS, Rackspace, etc.). That’s a vendor selection, not a framework for decision making.
- My strategy is IaaS, followed by PaaS, followed by SaaS. Cloud is not a logical progression of steps that you move through and implement.
- We are only doing private cloud. This severely limits your options and blinds you to the continuous innovations made by the cloud providers.
- We are only doing public cloud. Lifting/shifting the whole shop to the public cloud isn’t practical at this time. There are still workload types that simply don’t run well there.
- We are getting out of the data center business. This might be a valid data-center outsourcing statement, but it’s not a cloud strategy. Simply moving everything elsewhere to run does not make it cloud, it’s just traditional hosting.
The Strategy Document
In medium to larger organizations, it’s wise to have a clearly documented cloud strategy. It’s a living document that should be updated at least annually. Gartner suggests the following outline as a starting point for such a document:
- Executive Summary. A brief summary of the business drivers, challenges, and major steps involved in using cloud services.
- Cloud Computing Baseline. What mix of cloud services currently makes sense for the business? What services are industry peers using? What’s the current market analysis of the cloud providers?
- Business Baseline. What are the desired business outcomes (which should be measurable!) and the challenges/risks to the organization in adopting cloud?
- Cloud Strategy. What will you consume versus build? How will you secure, manage, and govern?
- Supporting Elements. What business processes and technical infrastructure need to be in place to use cloud effectively?
A good way to jump-start your strategy is to take advantage of free trials offered by the cloud providers. You can find out how the services work, where they shine, and where they are best avoided for your particular uses, all at no cost.
A good cloud strategy does not consist of a singular, perfect path. The strategies for Iaas, PaaS, and SaaS are all different. And within each of those strategies lie different paths, depending on the application or service in question. An all-encompassing, enterprise-wide strategy is likely to lead to trouble because you will end up trying to make everything fit into that one big idea that’s not optimal for anything.
The ultimate benefit of creating a cloud strategy is having a framework that you can use to guide your cloud decisions in a manner that gets the most value for your business. You will need to re-evaluate it from time to time as the cloud continues to evolve, but it’s a thing that you will use often.
Scott Kantner is the Chief Technology Officer of DSS, with expertise in IT infrastructure design and software engineering. Scott has been with the company since its inception, establishing corporate vision and selecting the technology that DSS delivers to its clients. Scott is an extremely knowledgeable IT consultant, with over 25 years of experience. He holds certifications from IBM, Tivoli, APC, Microsoft, and The Open Group.
Scott is responsible for DSS’s internal architecture, developing DSS’s customer infrastructure solutions, and working with customers to develop IT infrastructure solutions aligned with their business requirements. Scott’s experience spans multiple industry sectors, including banking, financial, manufacturing, energy, and retail. During the formative years of DSS, Scott participated in both the design and delivery of customer solutions and later became responsible for managing DSS’ systems engineering staff. As the company grew, the need to exploit new technology for both customer and DSS advantage also grew, and Scott became DSS’ first CTO. In addition to helping customers leverage their investments in technology, Scott is develops DSS’ portfolio of hosting solutions.
Scott holds a Bachelor’s degree in Computing and Information Sciences from Lehigh University. He is active in his church. Scott and his wife reside in Hamburg, PA, and have three children.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise, plus get instant access to more than 20 eBooks.