There’s no disputing that the big public cloud vendors want more of your data, your applications, your digital stuff to live in their respective clouds. But for most big customers, the decision is not about which cloud to go with but which of several cloud options best fits the job at hand.
This week Google (GOOG) announced HTC, the Taiwanese smart phone and tablet maker, as a big customer of Google Cloud Platform or GCP.
HTC is using Google Compute Engine, Cloud Google Cloud Storage and Google Data Store to build mobile apps that suck up less bandwidth and are less likely to freeze in dead-zone situations, according to a Google blog post by John Song, HTC’s senior director of cloud computing. He is one happy GCP customer.
But in a subsequent email exchange, Song noted that HTC corporate uses cloud services from Amazon(AMZN) and Microsoft(MSFT) as well. Why? Because some clouds suit some jobs better than others. Or the given department has a pre-existing relationship with the vendor, perhaps from the pre-cloud era. HTC’s information technology department, for example uses Microsoft Azure while HTC’s Connected Services unit is in the Amazon Web Services camp.
“When my group built our infrastructure, we were open-minded. We talked to Azure, AWS, SoftLayer/IBM(IBM), Ali Cloud in addition to Google,” he said via email. “In fact we were more inclined to stick with AWS as none of us had used GCP before and we knew AWS pretty well.”
In this case, Google won partly because of its data center replication story. Because Google owns its own fiber it is better able to control latency fluctuations between its data centers than the other vendors, Song said via e-mail. While big latency spikes and dips won’t necessary cause replication to fall apart, they can significantly ding performance, so that’s an issue.
He also credited Google for being the only one of the contending vendors to send an engineer for a pre-sales support call. The others all sent sales people.
This one customer-several clouds story jibes with what Sebastian Stadil, CEO of Scalr, a cloud management company, sees in the field. Projects within companies tend to be on either Amazon, Google, Microsoft or the OpenStack cloud framework available from many vendors, he said. But there are so many projects in large organizations that the organizations themselves end up being multi-cloud by necessity.
Recent research from Rightscale, another cloud management provider, said that among large companies (those with more than 1000 employees) surveyed, 10% use a single cloud, 82% use more than one cloud.
There are exceptions of course. With smaller companies in particular, there may be a need to keep vendor choices simple.
One San Francisco area company uses Amazon’s Redshift to analyze data flowing in from online customer interactions. Engineers at the company were interested in using Google BigQuery for some of that work. They tested it and loved what they saw but ultimately a business decision was made to keep everything in Redshift, according to one of those engineers. “Management didn’t want to introduce the complexity and overhead and security of managing yet another vendor,” he said, sounding a tad wistful.
This story was updated at 5:05 p.m. EDT with Rightscale data.