Post by vExpert Vladan Seget
Many small businesses that have reached peak utilization of their existing virtual infrastructure and need more storage, memory and CPU resources, are facing a dilemma: How can they expand their infrastructure?
VMware virtualized infrastructures are present in many organizations, both small and large, because of the technology’s reliability, stability and easy-to-use interface. I have never seen an admin wishing they could go back to managing physical servers once they’ve virtualized their environment with VMware.
Consolidating physical hosts was a good argument over the last 10-15 years, but when it comes to virtualizing everything - 100% virtual - there are few areas where you just can’t do it.
Usually SMBs aren’t fully virtualized yet because of several reasons:
- Physical Hardware Reliability – Over the last decade, hardware manufacturers have managed to provide hardware that’s more efficient, more silent and more resilient with less failures. Because the physical hardware installed years ago still works fine, it makes the decision to virtualize your server and applications sometimes problematic.
- Legacy Applications – There is still AS400 hardware that cannot be virtualized. There are applications not suitable for virtualization, like servers that use hardware USB dongles.
- Application Support – There are some vendors that still do not support applications in a virtual environment.
- Specific Hardware – Applications using modems, fax cards, etc. aren’t supported in vSphere or ESXi, so they must stay physical.
- Licensing Issues – Some vendors license in strange ways. They want to charge you for your whole VMware cluster even if the app is running on a single host.
- Performance – SMBs are almost always short on resources, making it difficult to invest in the best SAN device. Let’s say that you consolidate a workflow that has been running just fine on a physical host (1:1 ratio) to a host that now runs 10 VMs. Unfortunately, the recommended hardware wasn’t budgeted and instead, lower-performing hardware was acquired. You might see an I/O bottleneck – a bottleneck that did not exist when the app was running on single physical host.
When a virtual infrastructure has peak utilization and you start to lack resources like memory, storage, and CPU - what’s next? There are several areas worth exploring, and first of them is private cloud.
On-premises or Off-premises expansion?
The existing data center can always be expanded. Almost always. Even the smallest one. The topic of expansion is always a challenge for very small data centers, or data center rooms where there isn’t much physical room left with aging hardware, or where the existing hardware generates so much heat that the first question is, “Will our air conditioning manage to cool down the room?” However, there are always areas that can be leveraged to consolidate servers. P2V process for remaining physical hosts, for example, would be the first path to explore.
But a traditional approach for data center expansion would be to add more hosts to existing clusters or to create a new cluster with the same kind of hardware. Or adding a few more LUNs to spin up more VMs, etc.
So the on-premises expansion would involve adding more hardware to existing virtual infrastructure that would allow you to virtualize more applications and satisfy the growing needs for more storage, memory and CPU power.
Where SMBs might hit a wall is with budget. This is where it might make sense to run workloads in a public cloud off premises, like VMware vCloud Air. vCloud Air is an infrastructure as a service offering built on the foundation of VMware vSphere. You can migrate existing onsite virtual machines (VMs) or start up new application VMs directly in the cloud. You can also easily port VMs and other business-critical workloads back and forth to the location of your choice, all with the secure and capable foundation of vSphere.
Where cloud adoptions almost always start is with development and testing or non-critical apps. It’s always easy to start with a non-critical app, see how it goes, and then, once validated and tested, move on with more critical or enterprise-class apps.
One example can seen through dev-ops. Let’s say a developer needs to test workloads, new builds, and new versions of an app quite often. If the IT admin does not have capacity in the local data center, he can offer to move those workloads off-premises. The simplest way would be to create few instances in a public cloud and provide those instances to your developers so they can do their job. At the same time, the IT admin would have time to familiarize himself with the offering, the UI and management tools. In the case of vCloud Air, the transition would be seamless, as you don’t need special training or monitoring, and you can extend your data center to vCloud Air leveraging the same infrastructure, network, and management you use today. The offer is fully compatible with your on premises vSphere data center.
The expansion of virtual infrastructures requires those involved to decide what they need for the long term. There are numerous costs, constraints and options, vCloud Air provides SMBs with a cost-effective option with the reliability, security, and performance you expect, plus the same trusted VMware support you already use.
Ready to get started with vCloud Air ? Check out vCloud Air OnDemand and get instant, pay-as-you-go access to new infrastructure in minutes. Receive $300 in Service Credits when you sign up today.
Vladan SEGET is as an Independent consultant, professional blogger, vExpert 2009 - 2014, VCAP5-DCA/DCD, VCP 4/5. Vladan’s blog, ESX Virtualization, started as a simple bookmarking site, but quickly found a large following of readers and subscribers.