This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter's approach.
Hybrid cloud is the talk of IT, but to avoid costly, labor-intensive megaprojects you cannot escape, pay particular attention to minimizing implementation and management complexity. These questions will help you identify the best hybrid cloud architecture for your environment:
1. What are the top ways we will use our hybrid cloud in the next 12 to 18 months?
In the midmarket, the No. 1 answer is disaster recovery (DR). A secondary data center for DR is a luxury most companies can not afford. Now, public cloud services have put DR within reach of virtually all organizations. The key is to identify the enabling technology that minimizes complexity, maximizes automation and does not overtax the IT staff. Easy cloud DR solutions exist today for midsized shops; don't be lead into a heavy professional services project.
For larger enterprises looking to utilize hybrid cloud to optimize and free up expensive data centers, hybrid clouds are attainable, manageable options. For example, organizations using VMware might want to leverage Hyper-V because they're running so many Microsoft applications, while others want to leverage KVM for better flexibility, network outputs and drivers. For them, alternative hypervisors give them flexibility and significant cost savings. Public clouds, on the other hand, are a place where they can grow certain applications, conduct testing and development, and run non-critical applications. However, modern transformation technologies for cross-platform management are necessary to avoid monstrous and expensive system integration efforts.
2. Which public clouds do we want to leverage?
The public cloud space is continuously morphing, and that means there are many choices. End users may be quick to ask for the clouds they recognize: Amazon or maybe Azure. You, however, need to weigh all the factors, including price, scale, support and service.
There are public cloud self-service models offering attractive price points, but may not have the support staff if users have problems. On the other end of the spectrum, there are options delivering premium-level cloud and service packages for a substantial price premium. You need to analyze which cloud providers are best by asking the questions, "What do I get for this?" "What do I want to manage?" and "How hands-on do I want to be?" From there, consider the best possible options for management and migration. The answer will most likely be a mix of on-premise solutions, cloud solutions and services. Of course, mixing and matching can add significant management complexity if you are not careful.
3. Which on-premise platforms do we want to use?
Certain applications may require huge virtual machines (VMs), and the technical staff might find that only certain hypervisors can handle the requirements. Maybe another application needs high I/O, which will lead to a different platform choice. Cost versus performance is always a factor. No matter the particulars, you need to think about flexibility combined with ease of use and the least possible disruption. Companies want to manage their hybrid environments in the same way they manage their current environments: They want a single, comprehensive management platform. You must be able to seamlessly migrate workloads between hypervisors and maintain consistent, compliant management. It is very doable today, and thus we see the meteoric rise in hybrid deployments.
4. How will we manage the hybrid environment?
You need to consider compute, network and storage together and ensure their hybrid management construct is capable of easily spanning a range of on- and off-premise platforms and resources at a very granular level. To be operationally efficient, you need a single point of administration and management across the hybrid resource pool a self-service portal alone will not be sufficient for day-to-day administration operations.
The ideal would be to have a hybrid management solution that is lightweight, thorough and cost-effective. When managing a hybrid configuration, you need to base these choices on individual requirements: Are they highly regulated? Who will be accessing the information? What is the information? And so on. Regardless of the answers, the best situation is to manage the hybrid cloud and on-premise workloads from a single point. Optimally, the hybrid resources would just work seamlessly with the existing management portfolio.
5. How will this integrate with our existing operations?
Most people inherently resist change. One of the biggest challenges comes from the people inside a company, especially (and rightly so) those who are responsible for delivering a certain grade of service like IT staff. You have to be aware of this issue, and if adopting hybrid cloud means replacing familiar management consoles, retraining personnel and changing current workflows, employees will balk. Integration with existing operations is essential to successful deployments.
In addition, the hybrid project must scale. What works for a handful of technical people will not work when large scale production IT is involved and needs a very deliberate, orchestrated solution that is seamless with existing operations. Success of the project comes down to the ability to manage it. Integration with current tools and processes is key.
6. What skills will be needed to deploy, maintain and operate our hybrid environment?
IT staffers need to be able to analyze what they have today and what is needed tomorrow in terms of cost, performance, compliance and security, and then evaluate the choices. To do this, you need a strong working knowledge of both on- and off-premise management and integration. The hybrid cloud demands a shift in thinking. With on-premise infrastructure, IT teams had to do a lot of the physical underpinnings, such as hardware installation, wiring and networking, so those skill sets were heavily valued. The cloud takes away some of that and introduces a new application of those skill sets. Now, IT teams need to adapt their players and potentially hire new staffers with expertise in hybrid cloud functionalities, management, integration and administration.
7. How will we prevent vendor lock-in?
Before you ask this question, you should look around to see whether you're already locked in and don't realize it. Preventing lock-in requires vigilance against technical and financial constraints that could impede the very flexibility hybrid cloud is meant to create. Think about how lock-in occurs and make the hybrid choices that prevent it. For example, if you can seamlessly and easily move hybrid workloads between disparate platforms, that reduces lock-in. If you deploy a management solution that can span platforms, that also reduces lock-in. The other type of lock-in is long-term contracts. Vendors have endless incentives to contractually lock-in customers. With the speed of IT change and options, CIOs should be particularly wary of the multi-year enterprise license agreement (ELA).
Enterprises are well along the road to hybrid IT, but many are wary of getting stuck with hybrid cloud megaprojects. The best way to avoid that fate is to focus on flexibility, while leveraging current competencies and investments every step of the way. By minimizing complexity in the implementation stage and creating a flexible management environment that's intuitive for the operations staff, you can steer clear of unsuccessful, costly and labor-intensive hybrid cloud deployments. With today's hybrid technologies and solutions, no megaprojects are needed.
McLeod is the vice president of business development at HotLink. He brings more than 25 years of technology experience with a strong track record with both Fortune 500 leaders and start-ups. His broad background spans both information technology and telecommunications markets, and includes executive roles in sales, business development, product management and general management.