It’s a competitive world out there. In order to remain relevant to customers, every organization must continue to innovate while keeping expenses in check. IT departments are no stranger to these challenges.
Decreasing budgets and increasingly demanding customers are a few of the drivers forcing IT shops to change the way they use technology. These challenges are at the forefront of what is driving the interest in converged infrastructure (CI) solutions.
Converged infrastructure is the coming together of compute, storage, networking and software into one integrated solution. While the benefits of convergence are many, it can be summarized as providing increased agility, resilience and lower costs (both capital and operating) through efficiency.
Let me give you one example of this resilience. Let’s say that you have recently deployed a new FlexPod solution. (FlexPod is one of many popular CI solutions on the market today.)
One day your electrical utility experiences a major power failure. Previously you would be scrambling to shutdown critical IT systems before your uninterruptible power supply (UPS) batteries ran down. However, today you are calm and collected because CI solutions can be automated to move virtual machines (VMs) from a primary datacenter to a disaster recovery site or into the cloud with no interruption in service.
It is no secret that power grids in the US and throughout the world are increasingly stretched. As a result, power failures are not uncommon and the scenario above is more real than hypothetical.
With the resilience that a CI solution promises, you might be tempted to think that you can afford to scale back on power protection. It is certainly true that CI solutions are much more energy efficient than the systems they replace and this often leads many to believe that persisting with older, larger, existing UPS systems is a viable option. I want to illustrate two problems with this thinking and provide some best practices when powering a converged infrastructure solution.
Problem #1: UPS Efficiency
If you are considering server and network consolidation into an on-premises converged infrastructure solution, it is reasonable to expect that one of the major benefits will be a significant reduction in utility bills. If your UPS system has the capacity for an existing environment, it is likely that it will have the capacity for your new leaner environment as well.
What often gets overlooked is that the efficiency of a UPS system itself will vary widely based on the connected load. A UPS may be 98 percent efficient when close to its rated capacity, but it may only be 75 percent efficient at lower loads. While this variance in efficiency exists with all UPSs, it is even more pronounced in older devices.
This means if you choose to persist with your existing UPS and operate it at lower loads, you can still expect to see a reduction in energy costs due to fewer physical server, storage and networking components. However, due to the lower UPS efficiency at low loads, the overall savings on your utility bill may not be as pronounced as you originally expected.
If your existing UPS is also old or has not had its batteries replaced in some time, the added knowledge that it will also be less operationally efficient may be the last piece of evidence needed to consider a refresh. However, if that is not enough, consider problem number two.
Problem #2: Communication
In the example above, there was a beautiful simplicity about how the CI system seamlessly avoided disaster by automatically moving VMs out of harm’s way and keeping business running smoothly during a power outage. What I did not tell you was that the CI management platform cannot respond to a threat in the physical environment unless there is something in place to identify that threat and communicate it to the CI management software.
For me, this is the single biggest reason to thoroughly examine your physical environment prior to deploying any CI solution. You must have devices to monitor anomalies and threats in the physical layer and communicate them to your CI management software. Otherwise, all the talk about resilience will turn out to be an empty promise.
If you are determined to evaluate and prepare your environment for a converged infrastructure solution, there are some best practices to follow.
The first is to evaluate your existing UPSs to ensure that you have the capacity to support your new investment, the redundancy to ensure it remains available, and lastly, a system that will communicate utility power failures and UPS system issues such as battery failures or low runtime to the CI management software. There are products from APC, HP, Tripplite, Emrerson/Liebert, Eaton and others that meet the needs of converged infrastructure solutions. Eaton in particular also has some excellent reference drawings and videos on the topic here.
The second best practice is to ensure you have proper power distribution from your UPS. Many CI solutions involve blade servers which commonly require 208V power and use IEC C19 to C20 power cords. If you are not already equipped with 208V power distribution units (PDUs) this is another area that may require attention. PDUs can also be equipped to monitor power consumption down to the individual outlet, allowing you to monitor anomalies and power cycle individual systems.
Prior to failure, a computer power supply often draws more power. An intelligent PDU with monitoring at the outlet is able to catch this proactively by graphing typical usage and alerting IT staff to anomalies and a potential failure.
The last piece I want to highlight is the need for proper environmental monitoring. A high temperature condition can be just as detrimental to system performance and business continuity as a power failure. For this reason, an environmental monitoring solution is critical.
Many power manufacturers now offer intelligent PDUs that also provide the ability to add a temperature or humidity sensor to them. For more robust monitoring, that also includes leak detection and smoke detection as well as third party alarms. Keep in mind, things like a generator or stand-alone monitoring appliance may be more appropriate.
The goal is that all power and environmental events are logged and communicated to IT management as well as directly to the CI management software, which can then automatically take corrective action and move VMs.
The benefits of converged infrastructure are increasingly clear. And with a proper power protection and monitoring solution, you can be sure that your deployment will be a success.
For more on optimized power solutions, check out this article from BizTech