Last week, CenturyLink Cloud led a webinar focused on helping organizations kick-start their cloud efforts. While the cloud makes a lot of sense for organizations of all sizes, it can be daunting to get started with such a disruptive initiative. No doubt that most CIOs have heard horror stories about failed cloud initiatives, so how do they avoid becoming a statistic? CenturyLink Cloud offered five best practices to help companies take a successful plunge into the cloud.
#1 - Form a “tiger team” to achieve quick wins and pursue IT-as-a-Service
What’s a “tiger team”? I liked this description of its military background and applicability to IT:
“When the Navy needed some quick turnaround work or repairs, they would assemble a tiger team,” Ballard said. “The connotation was that it was a self-contained team that included all the skill sets and resources needed to do the work — journeymen, planners, engineers, fabricators, etc.”
“A tiger team was a small hand-picked, particularly skilled and capable group of ‘tigers,’ often chosen and chartered by a commanding officer, to plan for and/or achieve a very specific mission,” Lehman said.
The good news - you probably do this in your organization today when new, major initiatives come up and big organizational change happens. This cloud-oriented tiger team should be tasked with identifying strategic projects and executing them. Instead of pursuing complex, instantly-outdated, multi-year plans to embrace the cloud, start small and get moving! Ideally, this team is comprised of executive sponsors (from IT and other business units), a program lead, business analysts (who understand standard operating procedures and business needs), application architects, network engineers, and developers. This is a team of hands-on doers, not a bunch of box-drawing hand-wavers. Find the top performers in your organization and assemble them into a team that can partner with business units and start delivering high-value projects while aggregating valuable knowledge about cloud technology and IT-as-a-Service.
Another bit of good news – the individuals on the soon-to-be tiger team are already thinking of cloud on their own time, and have likely been itching for this kind of assignment.
#2 - Find the right cloud provider for your organization.
There are dozens upon dozens of interesting cloud infrastructure providers, and they aren’t all the same. Some are vanilla providers without much differentiation while others offer something unique.
The trick is finding the right one that is a match for your organizational culture, skills, priorities, or business needs. This isn’t unique to the cloud and applies to many (most?) of the strategic investments that a company makes. It’s likely that you have an evaluation process in place today that use assessment checklists, preferred vendor lists, strategic sourcing departments, and more. You also likely have a set of IT requirements that any new vendor needs to comply with. Here are some things to make sure you are thinking about:
*Security capabilities. What kind of physical, network, and account security does the provider offer? Are you running batch jobs with sensitive data, or hosting a public website with limited private data?
*Compliance certifications. Do you need assurances that the provider can host HIPAA or PCI-compliant applications? Are you storing data for EU users and weighing data sovereignty concerns?
*Performance. Does the provider offer server and network specifications that meet or exceed what you have in house? Does that matter for the types of applications you plan to run in the cloud? Can you provision components (virtual machines, storage, networks, etc.) at different performance levels to meet your needs?
*High availability. Are the various services you depend on individually designed to be highly available, and does the provider offer the necessary services (e.g. VPN) to architect highly available applications on the platform?
*Strong SLAs. Cloud failures are inevitable, so how does the provider define their SLA with regards to availability, response times, and outage credits?
*Disaster Recovery procedures. Do you have mature processes in place to manually fail over in the case of a full site outage, or are you looking for a provider who delivers a built-in DR capability?
*Global footprint. Location, location, location. Does the provider’s data center presence match your customer and employee geographies?
*Management interface. The creation of a cloud resource is invariably the smallest portion of its lifespan. So, it’s great if you can spin up individual servers or databases quickly, but how easily can you manage them after the fact? This may include installing updates, resizing capacity, changing security roles, or even shutting it down.
*Application services. Do you need more than raw infrastructure? Looking for managed databases or web platforms that abstract away the low level details? Or storage services that can be accessed from cloud – or on-premises – systems?
*Pricing plans. Are you looking for a fixed contract (or monthly commit) that results in lower costs? Or do you need a provider who delivers true pay-as-you-go services?
*Support. What level of support is needed? Do you want a dedicated support person? Call center? Public knowledge base? User forums?
*Professional services. Do you expect a consultative experience during the POC stage or when doing critical business updates? How about the full outsourcing of cloud server management to the provider?
Assess what really matters to YOUR organization, and find the provider who most closely matches those priorities.
#3 – Find, partner with, learn from, and collaborate with shadow IT.
What is Shadow IT? A phenomenon that occurs when there is unmet technology needs in an organization and employees “go rogue” and use non-standard solutions. You’ve likely read negative press about the concept and you might even have corporate initiatives meant to stamp it out.
Where do you often find shadow IT in your organization?
*Line of business applications. These are the applications used every day to collect, store, and analyze data used for business decisions. This can take the form of Microsoft Access databases, Filemaker Pro applications, Excel workbooks, or custom developed web applications.
*Productivity solutions. This is the technology that is meant to improve efficiency and collaboration. Bring-your-own-device (BYOD) brought this to the forefront as employees used tablets and smartphones at work, whether supported or not. Other things in this category include instant messaging and VOIP tools like Skype, and virtual meetingplace tools like Google+ Hangouts.
*Data management. Often departments need to store, analyze, or transfer data and don’t like the sanctioned options offered by corporate IT. This may mean creating production database replicas for later analysis (by custom BI tools!), or using storage services like DropBox to exchange files with business partners.
The bad news? These shadow solutions may not meet corporate security standards, are invisible to IT audits, don’t feed innovative ideas back to central IT, and frequently become too large and difficult to maintain locally.
Now, the good news – your employees are showing initiative and can show you exactly how IT can improve day-to-day productivity!
The one thing all the above categories has in common is how much EASIER each one is in the cloud. Today, you can use a credit card to provision enterprise-class CRM, ERP, marketing, and HR software in the cloud without batting an eye.
Instead of shaming and shutting down those in your organization who are trying to solve problems in new ways, partner with them and make them some of the most passionate advocates for your new cloud projects. Figure out what technologies they’ve pursued, how they evaluated them, and use their knowledge to make better cloud choices.
#4 - Learn and enable best practices in the cloud.
The cloud is very different from how you do on-premises enterprise IT. Resource planning, financial projections, technology choices, level of control, and self-service model are all very likely foreign to your IT department. Therefore, carefully consider how you discover and capture best practices in this new world. What are some areas to focus your attention on in order to truly take advantage of the cloud?
*Embrace automation everywhere. Automation is the difference between long-term cloud success and simply having a souped-up virtualized environment. Self-service is critical for the agility promised by the cloud, and that’s only possible if you’ve made a concerted effort to automate nearly every activity in the application lifecycle. This includes provisioning environments, updating environments, continuous application deployment, automated testing, auto scaling of servers, and much more. Could you immediately adopt an immutable server model where you never update a deployed server, but simply re-create it whenever there are application changes? If not, look at how you do these sorts of things internally today, and start introducing processes and tools that will make it that much easier to understand and exploit the cloud.
*Design software differently. New web applications (whether hosted in your data center or the public cloud) should be built for multi-tenancy, statelessness, horizontal scale, ready for failure, and service-oriented. This means building applications made up of components that can scale independently, communicate asynchronously, and have no dependency on a particular piece of hardware. Deploying an application in the cloud doesn’t guarantee high availability; rather, the application must be specifically architected in a way to be resilient from failures. This may be counter to how applications are built in your organizations today, but unless you design web applications in this manner, you’ll find it difficult to take full advantage of the cloud.
*Use the right service for the right situation. Not everything belongs in customer-managed virtual machine. Cloud computing is made up of many “as-a-service” offerings, some of which take the management burden away from the customer. Consider services like database-as-a-service where a provider manages the database infrastructure and the customer doesn’t have to worry about provisioning, scaling, backing up, or patching servers. Or object storage services that let you store massive amounts of highly available data without worrying about setting up that infrastructure yourself. If you don’t need complete control over the infrastructure, look to other types of cloud services that take that work away.
*Apply security everywhere. Get ready to secure your data at rest and in transit. Get familiar with data encryption concepts and tools, and don’t simply rely on peripheral security to protect your servers and data.
*Get your hands dirty! Maybe the most important item in the list. Every senior architect and developer in your organization should have a (free) account set up for all the major cloud providers. Try things out. Fail quickly and safely. Figure out what works and what doesn’t. The cloud has made it amazingly simple to explore and experiment with technology and you should be actively encouraging your technical team to try it out.
#5 – Carefully assess which applications to migrate.
Not every application is fit for the cloud. Many applications in your data center weren’t built with the cloud in mind. They may be monolithic applications that aren’t naturally redundant and resistant to failure. They could use hard-coded configurations and IP addresses, and expect physical proximity for all components. Simply moving these types of applications won’t grant magical performance or scalability improvements, and may actually result in a worse experience.
However, if you have cloud-friendly applications in your data center, you may realize value from moving them to a cloud environment. There are at least three techniques to follow.
First, you can *lift-and-shift and simply take the application as is and move it to the cloud.
Second, you can *refactor the application and make small changes to make it friendlier to cloud environments.
Finally, you could do an entire *rebuild of the application to completely optimize it for the cloud. Odds are, the last option is not economically beneficial, and it’s better to leave the application where it is.
The best migration candidates include vanilla packaged applications like Microsoft Exchange and Microsoft SharePoint, public-facing Internet websites, applications with bursty or unpredictable usage, applications/façade for mobile users, and custom service-oriented web applications. Consider leaving the following types of applications in your own data center: monolithic applications that can’t be easily broken apart, applications with extensive integrations with on-premises systems, heavy I/O applications or those optimized for specific hardware.
The cloud offers game-changing opportunities for organizations who take advantage of it. We’ve seen organizations both large and small take the plunge and realize the benefits. Interested in talking with us and getting some help on your cloud journey? Contact us today and we can help you avoid the pitfalls and get up and running quickly.