BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Selecting the proper hybrid cloud management tools is not easy since there's a level of hype surrounding them, and the tools available are still new and evolving rapidly. Other challenges include evaluating the tool's future potential, weighing it against future needs and avoiding vendor lock-in.
There are three strategic decisions buyers must make before wading into the technical issues.
The first decision: Single source vs. open source
Organizations must decide whether to buy a single-sourced bundle of hardware, cloud stacks and management tools, or take on the integration task themselves by adding open source or purchased stacks to management tools.
To make this decision, measure the progress of the hybrid cloud project. If it's just a prototyping effort, the decision will depend on the IT team's sophistication, as well as its schedule and budget constraints. Ideally, a prototype is a way to test a variety of approaches, which would make open source systems and tools not tied to particular hardware the better option. This would allow admins to evaluate a broader spectrum of vendors and approaches.
But if the project's budget is tight, schedule is short or is intended just to prove a concept, it might be best to go all-in with one vendor. Even in that instance, the intent for the next phase of the hybrid cloud project -- whether it is a broader evaluation or a move to production -- will impact your choice of tools.
The second decision: Hybrid cloud deployment model
Heading into 2017, hybrid cloud faces a fork in the road. The large cloud service providers -- Microsoft Azure, Amazon Web Services (AWS) and Google -- are heading toward offering a preconfigured private cloud software package. Microsoft has made announcements about Azure Stack, while AWS has a few installations, notably two major known U.S. government installations: CIA and G-Cloud.
Because these preconfigured private cloud software packages are new, the question of pricing is up in the air. Nonetheless, the idea of a single cloud stack that operates seamlessly across the whole hybrid cloud is appealing to admins. It's essentially the same operating model admins see between availability zones in the public clouds.
The downside is that the cloud providers have little incentive to interoperate with other providers' platforms, so hopes of hopping between public clouds, based on price or services, is likely on the back-burner for a bit. Microsoft has limited Azure Stack's availability to just one or two server models with only the major traditional vendors, so it isn't an unbundled software stack.
The third decision: Whether to outsource management
One alternative to managing your own hybrid cloud is to outsource that effort to a third-party vendor such as Rackspace. The result will likely be an OpenStack-based cloud with a mashup of current tools. It's a tradeoff between building in-house skills and the cost and limitations of an external vendor, but it can ease the hybrid cloud learning curve, especially around OpenStack.
Evaluation criteria for hybrid cloud management tools
Once a game plan is in place, and organizations address the three strategic decisions previously mentioned, the next step is to evaluate hybrid cloud management tools. The primary evaluation is between the complete management suites, and the specific tools that address only one or two areas.
For the a la carte approach to tool selection, evaluate each tool class separately, and ensure each can interoperate before making final decision.
Address the following questions when looking at prospective hybrid cloud management tools:
- Choice of base platforms. Is usage limited to selected vendors, with only one or two approved platforms, or will the tools manage any commercial off-the-shelf environment?
- Vendor lock-in. Does the tool require that everything in the platform be from the same vendor?
- Scalability of the tool. Can the tool manage your target cluster size? This could be a problem with pre-canned products such as Azure Stack, which currently carry restrictions to small cluster sizes.
- Public cloud or private cloud neutrality. Does the tool communicate with and manage a variety of public clouds? Does the tool communicate with and manage different private cloud platforms? Consider this, and the previous criteria depending on the rollout roadmap, since admins may not care to have a wide variety of clouds.
- Central management. Do all the tools come together at a single workstation? Are they accessible from one menu? Tool suites generally achieve this goal more easily than a mashup of different vendors' tools.
- Multi-use. Can multiple admins use the tools simultaneously?
- Functionality and performance. How well does the tool do its job? You might have a specific list of criteria to answer this question, depending on the tool. Test hybrid cloud management tools in a sandbox to identify the most fitting ones.
- The GUI. How good is the GUI? The quality of a GUI is crucial for a tool with a pivotal role in the new hybrid cloud.
- APIs. Does the tool have a standard API? If so, the apps that use the API have an open product that won't require recoding for another tool.
- Policies. Does it support a policy-based system that cloud tenants can use?
- Scripting. Does it support scripting? Are these scripts fluid across a variety of tools or clouds?
- Version control. Does the tool support versioning controls? Is the ability to control versions built-in?
- Cross-cluster synchronization. Does the tool support version synchronization across cluster nodes? It's usually important for code sets to be in sync across the cluster. Admins can achieve this using a separate tool, but each tool has to work in that environment.
- Product roadmap. Does the tool have a good path forward?
- Vendor reputation. What is the vendor's history with your operation, customer feedback, longevity and size?
- Cost. This is a major factor, and there's a wide price range for different hybrid cloud management tools. Remember that cost is a multiyear issue and, in a fast-evolving technology space, new add-on features become available frequently. Factor these costs into any software choice, especially if there are strong vendor lock-in risks.
The cost issue is even more pertinent if hardware restrictions exist, such as requiring original OEM add-on components like drives. Support costs can also be an issue. Watch out for large increases a few years after initial installation.
Classes of hybrid cloud management tools
Tools can be grouped as:
- Comprehensive tool suites that encompass most features;
- Orchestration software, perhaps just for servers, networks or storage;
- Software updates that keep nodes synchronized;
- Security tools that handle image validation, key management and more;
- Governance tools;
- Activity monitors, such as intrusion detectors, failure monitors and activity reporting tools;
- Billing tools to support a back-billing system;
- Policy template and management tools to allow for template creation; and
- Cost optimization tools that will optimize costs between public clouds and also determine private cloud placements.
While the purchasing process for hybrid cloud management software is a complex one, determining the type of tool an enterprise needs and how it meets specific criteria will help buyers narrow down the hyped-up options and choose the tool that will enable the most effective hybrid cloud management activity.
What a hybrid cloud management program does for companies today
What role can hybrid cloud management software play in an enterprise?