Top challenges in moving to a private cloud
There’s no question; cloud computing is essential to staying competitive in today’s economy. Whereas organizations were previously nervous about joining the cloud, it has been reported that 83% of enterprise workloads will operate in the cloud by 2020. Of that 83%, 41% of enterprise workloads will be run on public cloud platforms, while 20% are predicted to be private-cloud-based. But why is the private cloud base growing at such a slow rate?
The worldwide x86 server virtualization market has reached $5.6 billion with server virtualization rates in most enterprises crossing 75%. Yet, within those same enterprises, if you ask developers what they are using, the answer will most likely be Amazon Web Services (AWS). For some strange reason, enterprises are not deploying their workloads on virtualized servers.
Dominant virtualization players such as VMware have rolled out impressive cloud initiatives such as vRealize, yet, quizzically, their adoption has been low. And this isn’t just the case for vRealize, either. Other open source technologies such as OpenStack and CloudStack have experienced poor adoption rates in comparison to public cloud adoption rates.
In general, the enterprise IT world is stuck with traditional virtualization, while developers at the same enterprises have moved on to public clouds, even though having a private cloud can be more cost-effective. This article outlines the reasons why the mass switch to a private cloud has not yet happened, including legacy structure, labor costs, and private cloud risks.
Legacy Hardware Infrastructure
A primary reason why enterprises tend to be averse to private clouds is due to legacy hardware infrastructure. More often than not they are unable to utilize their existing infrastructure. Other than the high costs involved, rolling out new infrastructure means juggling multiple vendors for virtualization, storage, networking, etc., leading to vendor lock-in and increased management complexity.
Money and Time Constraints
In a bid to avoid vendor lock-in, many enterprises turn to CloudStack, an open source cloud management platform for delivering Infrastructure-as-a-Service (IaaS) in cloud computing environments. In essence, CloudStack is a one-stop shop for creating private clouds. Unfortunately, it takes at minimum a couple of weeks to deploy using CloudStack–entirely too long. With Cloud Stack, it is difficult to scale up, leaving it viable to deliver a subpar user experience.
Additionally, public cloud vendors have a skilled workforce on their payroll. Does your IT staff have an equal level of skill and expertise? CloudStack needs various expert cloud engineers, and those engineers do not come cheap. In fact, expert manpower accounts for a grand majority of the operating costs of a cloud, more than servers, storage, or networking. Below are five roles or teams that should be added to the cloud maintenance team.
- Hardware procurement teams
- Data Center design and build teams
- Operations staff is needed around the clock
- Security personnel during the design, build, and operations process.
- Networking teams support highly available networks. Their measured knowledge helps to design, debug, scale and operate the network.
To hire and financially support these teams, companies must invest heavily in an area that is not their core offering. In face of such complexities, enterprises prefer to choose public clouds such as Amazon where creating a virtual machine is as easy as 1,2,3. The supporting steps may be complicated, but AWS makes allocating storage and attaching it to a virtual machine quite simple.
Private Cloud Risks
There are many perceived and proven private cloud risks:
Without the above-mentioned experts on the payroll, everything that happens on your VM is your responsibility. That means a private cloud could be less secure than a public cloud.
In the public cloud, companies pay a premium to have a steady level of performance output. It is outlined in their SLAs. Yet, with a private cloud, it can be difficult to predict many performance issues on the horizon, for instance, how changing loads will affect application performance and user experience, or how the hardware that you’ve hand-selected will actually work together at an optimal level.
Private cloud is not elastic; enterprises are limited by the hardware they have at hand. If the workload needs expanding and you run out of hardware, you have no choice but to procure new equipment, make space for it, rack it, and get it up and running, which can take an unreasonable amount of time.
These are some fundamental challenges and private cloud risks that make it difficult for cloud projects to get off the ground. In our next post, we discuss more issues such as compliance, security challenges, user experience, and disaster recovery that prey on the minds of enterprises when they consider the shift to a private cloud.
Learn how Accelerite acquires Convirture for Cloud Management here.