Clouding the Enterprise Definition: Verizon Announces Potential 48 hour Outage
It what many have called a surprising move, cloud provider Verizon has announced to some customers that it will be experiencing a service outage of up to 48 hours. While this doesn’t surprise some who understand the nature of cloud services and the service level agreements wrapped around them, there are concerns that this is affecting what Verizon themselves call an enterprise-class cloud environment.
During times when the cloud wars continue to heat up among service providers in Tier-1 to Tier-3 cloud environments, many customers are finding themselves left out in the cold when these sort of incidents occur. There is also some genuine reasoning for the upgrades that often target security and resiliency upgrades. Many are also noting that this does not violate the service agreement for the provider as signed by customers of the Verizon services.
This is a great conversation starter for us in the industry around what we are able to call an enterprise-class or carrier-grade system. These monikers should not be used lightly, and can be particularly challenging because they have different meaning to different customers.
Imagine a customer who is using the enterprise-class service with a distributed N-Tier application scaled across multiple regions. For that customer, this may only be a mild irritation. Does this mean that we are able to say that the customer has to be prepared to be enterprise-class with their application architecture to truly enjoy the merits of a cloud SLA?
Here be Dragons…I mean outages
There is a strong lesson here: customers hold the final responsibility for application resiliency. This is a hard lesson learned for many organizations who run to the cloud with hopes that resiliency is built-in. We are seeing more situations where vulnerabilities require rolling reboots, upgrades, and more among both on-premises data centers, and public cloud environments.
Being prepared with an application architecture that leverages the scale-out capability of distributed cloud environments is the real path to achieving enterprise-class protection. Operations teams in your on-premises data center will tell you their past results, but as financial advisors say in every prospectus: Past results are not indicative of future performance.
Ben Kepes did a nice write-up on the situation over at Forbes, and hopefully we will see some good lessons learned by both providers and customers on how outages can affect our cloud environments.