Verizon has promised users of its cloud service that last weekend's extended period of downtime will be the last they're subjected to, thanks to the rollout of a new update.
The company has souped-up its Verizon Cloud offering to include new "seamless upgrade" functionality that should let Verizon push out future system upgrades without any service interruption.
"Traditionally, updates have been made via rolling maintenance and other methods," said Verizon in a blog post.
"Many cloud vendors require customers to set up virtual machines in multiple zones or upgrade domains, which can increase the cost and complexity.
"Additionally, those customers must reboot their virtual machines after maintenance has occurred," it added.
The Verizon Cloud upgrade means this fate will no longer await its users, as the company claims future upgrades to the service will be able to happen in the background, with no knock-on effect for customers.
Outage outrage
The news that some of Verizon’s enterprise cloud services could be taken offline for up to 48 hours last weekend prompted an outpouring of dismay from analysts and end users.
In a statement to Cloud Pro, the company stressed the downtime would only affect users of its Verizon Cloud public cloud platform.
“We are conducting a round of scheduled maintenance for Verizon Cloud beginning on Saturday 10 January at 1am EST. This maintenance will affect production nodes limited to Verizon Cloud and will only impact Verizon Cloud customers.
“Customers using Verizon’s legacy platforms (Enterprise Cloud, Enterprise Cloud Managed Edition, and Enterprise Cloud Federal Edition) will be unaffected,” the statement read.
Furthermore, while it allowed a maximum of 48 hours for the maintenance to take place, it said at the time the work was likely to be done and dusted before then.
“We have informed clients to be prepared for the system to be down for up to 48 hours, although we do not anticipate the work taking that long,” the statement continued.
Even so, this didn't stop the company coming under fire from end users on social networking sites and from market watchers for even daring to suggest its service could be offline for such a lengthy period of time.
In the end, the system upgrade was reportedly completed by around 5pm ET on Sunday.
Feeding back
Speaking to Cloud Pro ahead of last week's downtime, Clive Longbottom, service director at IT analyst house Quocirca, said even though the work was taking place at the weekend, it could still harm business productivity.
“Forty-eight hours is the sort of time that would be seen as being necessary for a complete forklift upgrade and it makes no odds whether it is at the weekend or staggered over nights, as these are now meaningless on the global stage,” he said.
The projected length of the upgrade was also completely unnecessary, Longbottom continued, and suggests shortcomings in the way the Verizon Cloud is built.
The platform was launched in October 2013 and offers users access to public cloud-based compute and storage services, and draws on the communication provider’s myriad of datacentres dotted around the world and its networking capabilities to deliver them.
“It shows an insane lack of understanding from Verizon on how a cloud should operate,” Longbottom said.
“If a cloud is architected correctly then it can be taken down piecemeal for maintenance, still providing adequate access to the functional parts of [it].
“Verizon should have architected its cloud correctly: it should be able to take components out of the cloud and replace them on a bit-by-bit level without impacting the cloud itself,” he added.
However, James Staten, vice president and principal analyst covering infrastructure and operations professionals at fellow market watcher Forrester, told Cloud Pro the outcry prompted by the downtime was out of proportion.
“It only affects [its] newest Verizon Cloud service, so it does not affect any of its other cloud or managed hosting services. And, given how new this service is, there really aren’t all that many customers or workloads on [it] yet.
“We estimate that less than 10 per cent of [Verizon’s] cloud/hosting customers have deployed anything to the new Verizon Cloud yet,” he added.
Additionally, the downtime is only scheduled to affect one out of the three compute resources (which include Reserved, Elastic, and Research) currently available on Verizon Cloud.
“Only Reserved is affected by this outage. Granted it is the primary resource type but most customers have not committed to these resources yet,” said Staten.
“As you would expect, companies tend to test things out on Research before deploying to Reserved, just as you would by building the app in test and development, then moving it later to production.”
He also echoed Verizon's claim that last weekend's outage will be the last round of maintenance for Verizon Cloud that would take that long.
“There are some elements of the architecture that were not optimised at first release. One such issue is not having the ability to apply firmware changes non-disruptively,” said Staten.
“This outage will add this capability and, as such, the hope is that this will be the last long full-service outage customers experience.”
Long-term effect
In light of the fact one of the main selling points of cloud is its “always-on” status, Longbottom fears that – even if this is the last time Verizon embarks on such a long period of maintenance – end users could be put off using public cloud by episodes like this.
“This will give public cloud a bad name – and will need strong messaging from the competitors out there to be able to ensure that people understand that this isn’t how cloud needs to be,” he explained.
“It’s bad enough with unplanned outages in the likes of AWS and [Microsoft] Azure – but I just cannot understand how Verizon is messing this up to such a major extent.”
