Many cloud naysayers believe that cloud provider service outages, like VM security concerns, are a major problem....
But this argument is overblown, because there are ways to protect against intermittent downtime.
These pessimists use the same, tired arguments to rail against the cloud (usually some variation of "When my provider goes down, there's nothing I can do!"). These feelings stem from a lack of control over the situation, because the infrastructure is outsourced to a cloud provider. But are they warranted?
As a fun exercise, let's apply the same logic to common, everyday problems:
- When the power goes out, I have little recourse for the loss of service.
- When my flight is delayed because of the weather, I have little recourse for the loss of service.
Those situations don't make national news because they are a part of living in today's society. Furthermore, smart consumers protect themselves against these possibilities, if they find the losses unacceptable. For instance, you can protect against a power outage with battery backup units and diesel generators. You can also fly standby to get to your destination.
Service outages occur, no matter who's in charge
Cloud naysayers assert that virtual machine (VM) downtime caused by a cloud provider is far worse than instances when internal IT causes a service outage. Regardless of their reasons, the feeling is irrational. At the end of the day, a VM that is down is still a VM that is down. No matter who caused the service outage -- whether it was a cloud provider or internal IT -- you rarely know exactly when a VM will be operational again.
To counter proponents of in-house infrastructures, compare a cloud provider's level of experience hosting VMs to your own.Simply put, people who host VMs are in the business of hosting VMs. Contrast that position with your IT staff's expertise. When a cloud provider goes down, its workers are more likely to have the tools and processes in place to quickly resolve the issues. You may have an experienced virtualization staff, but cloud providers have workers who eat, breathe and live VM hosting.
A smart approach to handing cloud provider service outages
Moving VMs to a cloud provider requires a smart approach. The Amazon Elastic Compute Cloud, for example, can host VMs across multiple data centers in various locations, which provides a level of redundancy in case of a data center service outage. Amazon's infrastructure and usage-based billing also allows IT departments to build customized high-availability infrastructures as a site failure contingency.
Even today's hypervisor vendors offer products that can manage locally hosted VMs alongside workloads that reside with one or more cloud providers. (VMware vCloud Director, for example, provides this single-pane-of-glass management.) Now, bridging local VM resources with a cloud provider requires little more than integrating a few products.
Some argue that our industry's shift toward VMs that run elsewhere is an evolution of trust. It wasn't long ago that the idea of running IT services atop a virtual server was laughable. In a very short period of time, virtualization has evolved into a best practice. One assumes that a similar shift toward cloud workloads will occur as we slowly develop trust in hosting providers.
Addressing and improving cloud performance issues
Cloud computing outages: What can we learn?