Mathias Rosenthal - Fotolia

News Stay informed about the latest enterprise technology news and product updates.

Google cloud continues enterprise push amid more disruption

Google Cloud Platform kicked off August by adding security features, dropping prices and buying an ISV-focused startup, but networking continues to be a bugaboo.

It's been a busy start to the month for Google, turning to some old tricks (price cuts) and newer ones (enterprise features) to boost the status of its cloud platform -- and sandwiched in between was yet another service disruption.

The string of moves started last Monday, Aug. 1, with the general availability of customer-supplied encryption keys for Google Compute Engine. The enterprise check-list item followed a week later with the acquisition of Orbitera Inc., a startup focused on cloud-based commerce for ISVs. Then, on Tuesday, Aug. 9, Google cut prices on its already discounted Preemptible VMs by up to 33%, harkening back to a couple years ago when it tried to drive the cloud conversation by being the lowest priced option on the market.

Google already encrypts all data at rest in Compute Engine, but this new security feature gives companies even more control. Workiva Inc., a financial services software provider in Ames, Iowa and Google Cloud Platform customer, worked closely with Google to drive the requirements around the bring-your-own-keys capability, said Dave Tucker, Workiva vice president of engineering.

"Our customers want to know more about how you do encryption and how you do keys," Tucker said. "Everyone wants to supply your own keys for any data you put out there."

The ability to bring your own keys is a basic functionality to operate a public cloud, which Amazon has offered for years, said Carl Brooks, an analyst at 451 Research. The average customer doesn't care about where the keys are -- but it matters for enterprise clients with more complex networks or organizations with strict regulatory or compliance demands.

Orbitera deal another move to the enterprise

Google has acquired a handful of smaller startups to fold into its cloud platform since Diane Greene took over the unit last November, and Orbitera fits that bill. The West Hollywood, Calif., company automates commerce components for cloud services for ISVs, service providers and channel organizations. More than 60,000 enterprise stacks have been launched on the platform, according to Google.

It's unclear exactly how Google will integrate the service into its platform, but one of the knocks it has faced is a lack of partners and integrations, so this could help fill that void. Currently, Orbitera customers can use its service on AWS and Azure, with support from Rackspace and CenturyLink coming soon, according to the company's website. Google wasn't even on the radar, so this at least gets the company in the conversation to build its enterprise support, said Meaghan McGrath, an analyst at Technology Business Research, in Hampton, N.H.

Orbitera will maintain its neutrality in supporting multi-cloud commerce, Google said, and this fits with other moves it has made recently, including the beta rollout of Stackdriver, which enables multi-cloud management, McGrath said. It's also an acknowledgement that Google competes in a market where most customers already use AWS, which means it needs to make its cloud work alongside those implementations.

"If Google cloud started from the very beginning, they'd probably want all that business, but at this point they're pigeonholed into coexisting with AWS and, to a lesser extent, with Azure," McGrath said.

Another networking failure

Both of these moves are viewed by industry observers as positive steps, if not major leaps, for Google to gain ground on the long lead of Amazon Web Services and even Microsoft Azure. But the momentum was marred slightly by a widespread service disruption last week.

The problem started at 12:55 a.m. Pacific Time, Friday, Aug. 5, with some customers experiencing network latency and packet loss to Google Compute Engine, Cloud Virtual Private Network, Cloud Router and Cloud SQL until engineers resolved the compounding problems caused by a faulty router, according to a postmortem from Google.

The disruption lasted 99 minutes and all regions were impacted by the disruption, though the U.S.-West region was only down for 10 minutes, according to uptime tracking company CloudHarmony.

Workiva noticed disruption for some of its clients, Tucker said. The company distributes workloads to prepare for these types of incidents, and affected customers were notified once the problem was spotted.

"People understand that things like this can happen, but more surprisingly it's been so stable for us lately that it was really more of an anomaly than it's been in the past," Tucker said. "They're getting stronger all the time and we're very comfortable with the overall sustainability of the platform today."

Customers are often better equipped to deal with disruptions today than they were in the earliest days of public cloud, said Melanie Posey, research vice president at IDC, in Framingham, Mass.

"I remember seven or eight years ago having these calls in response to an AWS outage or Gmail being down and everybody would start freaking out and saying public cloud is unreliable," Posey said. "This seems like it was a networking thing, and networking things can happen in the public cloud."

Networking has been a problem for Google, and in April it had a global outage due to a network failure. This latest incident had less of an impact, partly because of the time it occurred, but also because it only affected internet control message protocol checks, while the transmission core protocol traffic remained functional, said Jason Read, founder of CloudHarmony, which is owned by analyst firm Gartner.

Regardless, it isn't a good sign for Google, and the relative quiet around the disruption highlights how far it is behind Amazon, Read said.

"Had this been a global [Amazon] EC2 outage, which has never occurred in my years of monitoring this stuff, it would have been an apocalyptic event and headline news for days," Read said.

Price drops back again

Google rounded out its latest set of moves with a price drop to its Preemptible VMs. The service was rolled out last year and provides cut-rate instances for batch jobs and fault-tolerant workloads used for short periods.

Cycle Computing, which provides orchestration software for big compute and big data projects, works with AWS, Azure and Google Cloud Platform. It's seen three times growth over the past year and increased interest in all three platforms, particularly Google, said CEO Jason Stowe.

"We anticipate this is going to accelerate that evaluation even further for financial services, life sciences, manufacturing customers -- their workloads tend to be friendly toward preemptible instances," he said.

While price remains an important consideration, price drops don't seem to resonate as they did two years ago when every tit-for-tat move from the hyper-scale providers generated headlines. Many cloud customers now find themselves spending more to take advantage of higher-level services, or better understand how to control their own costs.

"We're at a point [of] how we do optimization, in terms of estimating usage and prepaying and things like that, where we've got a pretty stable model with how we work with them on a pricing standpoint," Tucker said. "We always love to see price cuts from anybody, but it's just not like it was there for a while."

Trevor Jones is a news writer with TechTarget's Data Center and Virtualization media group. Contact him at [email protected].

Next Steps

Evaluate container services from AWS vs. Azure vs. Google

Google rolls out new cloud security features

Spotify goes all in with Google big data services

Dig Deeper on Open source cloud computing