This report looks at the extensive scheduled downtime announced this week by one of our largest and most well-recognized competitors:

• Are You Freaking Kidding Me?
• Hey, Where’s the Cloud?
• Why So Extended?
• What About Failover?
• Know Your SLA
• Don’t Run for the Hills

Are You Freaking Kidding Me?

No one was very impressed with Verizon’s announcement that its enterprise system Verizon Cloud would go down early Saturday morning for up to 48 hours. When Sharon Gaudin of Computerworld reported that the service would be moved offline beginning at 1 AM EST on Saturday, January 10, readers were quick to express their disgust.

One user, 99BitterReality, thought that the move by Verizon should be seen as its swansong, insulting the intelligence of anyone who would give them a second chance: “Only the truly stupid would continue with a company unable/unwilling to provide even a really lousy level of service.”

Commenter vipbackchannel noted that Verizon was responsible for the Affordable Care Act debacle in late 2013, when the system became widely unreliable right at the time when the Obama administration needed it most. In closing, vipbackchannel summarized his opinion of the telecom giant’s infrastructure: “They have very bad design.”

Hey, Where’s the Cloud?

As the title of this piece suggests, users of the Verizon service were not given any real options but instead had to frantically try to prepare for an outage of up to two days.

In an email to Yevgeniy Sverdlik of Data Center Knowledge, though, Verizon representative Kevin King did say that the maintenance process would probably not take up the full scheduled window.

Cloud downtime is not unheard-of. It was just a few months ago – September – that an unscheduled XenServer update resulted in reboots at major clouds, including those of IBM, Rackspace, and AWS.

Verizon introduced Verizon Cloud, which uses a different platform from its other enterprise cloud systems, in 2013. The shutdown over the weekend is not impacting users of any of the other, first-generation Enterprise Cloud systems. However, Sharon reported that the company has been trying to move its customers from those legacy environments to Verizon Cloud. A spokesperson for the company said that approximately 1 in 10 enterprise users were subject to the outage.

King said that the shutdown would allow Verizon to update its platforms at its data centers in Culpeper, Virginia, and elsewhere. He made it sound like it wasn’t a big thing: “Updates of this nature typically require some system downtime, and we notified customers in advance, so they could plan accordingly.”

Why So Extended?

Is this really newsworthy? Well, yeah, it kinda is. Yevgeniy notes that both the length and breadth of this shutdown makes it remarkable.

Bill Peldzus of analyst group Cloud Technology Partners said that the amount of time set aside by Verizon is substantial, even if it gives the company some padding.

Luckily, not every Verizon enterprise user is relying completely on them. Clearly they can’t handle that responsibility – this announcement seems to demonstrate that loudly and clearly. Some customers have a second cloud integrated as a backup in the event of disaster recovery. For those who don’t, Bill noted that this event could prompt them to reconsider their IT layouts.

What About Failover?

Public cloud providers typically have failover mechanisms that allow their users to transition their resource needs to a different datacenter when downtime is needed for one architecture. Peldzus commented that it was extraordinary for a hosting service to just warn customers their virtual machines would be unavailable and not point them toward a temporary, secondary system.

Know Your SLA

Everyone involved in this predicament must be aware of their service level agreement (SLA) with Verizon. Sverdlik says that some agreements probably allow a small percentage of downtime each year.

Industry analyst Peter Roosakos of Foghorn Consulting said that knowing outage safeguards is essential to choosing the right service: “It is up to the cloud user to understand the SLA of their cloud provider and design [or] plan accordingly.”

If your company has applications in place that must be live 24/7, you can stay live within one cloud if the hosting service uses “rolling maintenance.” In that scenario, the provider shuts down a zone at a time, so you can switch to a zone that is still up throughout maintenance. Mission-critical applications can alternately be backed up with the use of a second provider.

Although it should be noted that Verizon is excellent at the limbo, Peter recommended that users keep their expectations low. “As you’re implementing your applications on [outside systems],” he said, “you should always plan for and architect for failure.”

Don’t Run for the Hills

If this whole Verizon mess is making you feel sick, brother, don’t run for the hills. Run to us. We have got you covered, with a 100% network uptime SLA.

By Kent Roberts