06 Oct Overcoming Inertia — Mitigating Hyperconvergence’s Perceived Challenges
In the 2015 State of Hyperconverged Infrastructure market report, IT pros identified a number of reasons why they are reluctant to move forward with hyperconverged infrastructure initiatives (Figure 1). Many of these reasons were exposed in the State of Hyperconverged Infrastructure report, but there are also other concerns that people have around hyperconvergence, particularly when it comes to potential job impact.
Figure 1: Primary Reasons for Not Deploying Hyperconverged Infrastructure
Not everyone is enamored with the potential for hyperconverged infrastructure. Some simply have no present need to re-examine data center operations. There are myriad reasons why respondents identified these challenges around hyperconverged infrastructure adoption. However, the top reasons why people aren’t yet looking at the technology have nothing to do with the technology itself but rather have to do with the business cycle:
- Current solution works just fine — The adage “If it’s not broke, why fix it?” holds true for many. However, that won’t always be the case. Business priorities change on a dime, and understanding the significant benefits that may come from modernizing the data center with hyperconverged infrastructure solutions will be useful for the future.
- Recently upgraded infrastructure — For most hyperconverged infrastructure vendors, there is no need for a forklift. Different solutions offer varying degrees of integration opportunities but most can integrate into the existing environment at some level. Whether new applications are being deployed or there is a specific use case (such as VDI or ROBO), there may be an opportunity to introduce hyperconverged infrastructure into the environment.
- No current IT or business need — Some people truly have no present infrastructure needs and are focusing their efforts in other areas. Again, this won’t always be the case though, and these folks can sometimes be considered as part of the “current solution works just fine” category.
Here we’re going to tackle some of the more serious issues that people have brought up with regard to hyperconvergence.
The IT Staffing Challenge
We hate to be bearers of bad news, but most companies want to limit the number of IT staff that they hire, at least in certain areas. Infrastructure is almost always one of those areas because it’s directly associated with the expense side of the ledger. As a result, many companies are not willing to scale their infrastructure staff at the same rate that they scale the infrastructure itself. As a result, the IT staff is forced to take on more and more responsibility while working with dwindling resources. This is the very definition of that loathed phrase, “doing more with less.”
This situation is one of the very reasons why people consider hyperconverged infrastructure. By massively reducing the variety of hardware and software that needs to be managed and maintained, the same number of IT pros can manage an increasing number of appliances providing all of the necessary services.
Hyperconverged Infrastructure and Your Job
Job security is a huge concern for many IT pros, particularly when waves of outsourcing have taken their toll. At first glance to many, the words simplicity and efficiency may appear threatening, but there are different ways to think about this.
Those most impacted by hyperconverged infrastructure are likely to be storage experts, because hyperconvergence pretty handily disrupts this area. Remember that many companies believe that storage is the most expensive and complex construct in the data center. Further, many storage pros end up spending a whole lot of time on relatively mundane tasks. If you’re a storage expert, your company is probably hoping not to hire more of you anytime soon.
The goal here is to ensure that the services are still provided. That may not mean building out a SAN environment anymore. It may mean that the storage team should take the lead in helping to choose the right hyperconverged infrastructure solution — one that has the storage characteristics the organization needs. From there, if you are a storage pro, it will mean expanding beyond the storage realm into the world of servers and the hypervisor.
Hyperconverged infrastructure is the new animal in the data center jungle, but it’s tearing through it in a big way. The overall packaging of hyperconverged infrastructure is relatively new. That is, packing everything into appliances and stacking them together is new, but the internal components aren’t, depending on the solution.
There’s a simple way to ensure that your intended solution can meet your needs: test the heck out of it.
Every hyperconverged infrastructure vendor on the planet will allow you to do proof-of-concept testing of their solution.
Your job is very clear. Test it. Put it through its paces and decide for yourself if the solution will meet your needs. Use real-world tests, not synthetic benchmarking tools to perform testing. After all, you won’t be running synthetic tests in production; you’ll be running real applications.
Eliminating the Modern Refresh Cycle
The data center refresh cycle is out of control and is way too tied to capital budget dollars. Many organizations tackle individual resources separately. For example, they replace storage in Year 1 and then every four years thereafter. They replace a few servers each year. They replace supporting equipment, such as WAN accelerators and SSD caches, every three years. This mish-mash of replacement cycles creates confusion and doesn’t always result in an equal distribution of budget dollars.
What if you could implement a real rolling refresh process and simply add new appliances as new business needs dictate? And, as a reasonable replacement cycle emerges, you simply remove an appliance a year and cycle in a new one. Figure 2 gives you a look at once such scenario.
Figure 2: Hyperconvergence Refresh Process
The previous figure also demonstrates how you scale this environment. As you need to add new appliances, you just add new appliances. In the example shown in Figure 2, you start with three appliances in Year 1 and scale to four appliances in Year 2 and Year 3. In Year 4, you maintain four appliances but begin to cycle new appliances in. The entire process consists of racking and stacking new appliances as well as removing old ones.
This is a pretty new way to handle every infrastructure element. You no longer have to juggle storage, servers, and other supporting systems, including WAN accelerators, backup storage, and the other myriad appliances that litter the data center.
Obviously, though, you can’t just tear everything out of the data center today and replace it all with hyperconverged infrastructure. Like most things, you need to be able to phase in hyperconvergence as it is possible from a budgetary perspective. There are a couple of ways you can do this:
- Introductory project. If you have a new need, such as VDI or a ROBO modernization initiative, you can introduce hyperconvergence by implementing your new project on hyperconverged infrastructure. From there, on your natural replacement cycles, begin to migrate your other workloads to the new environment. Eventually, you will have fully phased out your traditional storage and server environment.
- Make coexistence a priority feature. Some hyperconverged infrastructure solutions enable peaceful coexistence with legacy infrastructure. For example, you’re able to use existing vSphere hosts with the storage in the hyperconverged cluster. Coexistence with legacy infrastructure is not possible with all hyperconverged solutions, so choose carefully. Some solutions require you to implement hyperconvergence as a standalone silo of infrastructure.
Understanding how to overcome inertia in order to make the right changes in your data center is really important. Also important are the economics behind the intended solution.