ra2 studio - Fotolia
SAN FRANCISCO -- Google has staked its claim in the cloud market on higher-level services and next-generation workloads, and the strategy resonates with IT pros here.
After a day of emphasis on mainstream enterprise case studies at its Google Cloud Next 2017 conference this week, Google rolled out a slew of new services to differentiate from its more successful public cloud competitors. Among the updates were an expansion of its popular ad hoc search tool, lower prices and discounts for long-term use, more security features, and expanded serverless computing and database options.
One appealing aspect about Google is it's viewed as a platform versus just infrastructure, said James Raby, project manager at Lonely Planet, a travel publisher based in Franklin, Tenn. "You don't want to be messing around with machines and networking and configuring," he said.
Regarding serverless, users can now bring frameworks to Google App Engine on top of the seven languages it already supports. And Google Cloud Functions is now finally available in public beta after a lengthy alpha window.
Google Cloud Functions has some ground to make up on Amazon Lambda, Amazon's serverless tool that debuted in 2014 and has been integrated into various Amazon Web Services (AWS) offerings. Google Cloud Functions currently only supports Node.js, though one of the more interesting capabilities is its integration with mobile developer service Firebase.
BigQuery has been the biggest differentiator for Google Cloud Platform, so it's no surprise Google continues to expand the capabilities of the managed service. Google added pipelines to its popular advertising services, such as Adwords, DoubleClick and YouTube Analytics. BigQuery can also now link to licensed commercial data sets without any additional code through a new program called Commercial Datasets that so far includes the collections from Xignite, HouseCanary, Remine, AccuWeather and Dow Jones.
Within Google Cloud Platform, BigQuery can also now search data from Google's NoSQL database service, Cloud Bigtable. Cloud Dataprep, a new service that also integrates with BigQuery and other Google Cloud Platform services, detects schemas, types, joins and anomalies, and it creates visualizations of data and suggestions based on usage patterns.
And following its partnership with SAP, Google added PostgreSQL support to its fully managed Cloud SQL. PostgreSQL is an important option for enterprises that want the closest open source equivalent to Oracle's database. The move follows AWS' addition of PostreSQL support and follows a larger trend by cloud providers to check all the boxes for enterprises that remain wary of moving their databases off premises.
Cloud Spanner and machine learning stand out
Amazon maintains a healthy lead in terms of market share, mind share and raw features, but taken collectively, Google has started to assemble a competitive offering to appeal to the enterprise market, said David Smith, an analyst at Gartner.
Cloud Spanner and the Video Intelligence API were the most interesting new services discussed at the conference, Smith said. Though not alone in moving in this direction, Google's emphasis on abstraction gets back to the original promise of the cloud, he added.
David Smithanalyst, Gartner
"In a lot of ways, whenever anyone is talking about serverless, whether that's Google or anyone else, they're talking about the original cloud vision that has somehow gotten watered down over time," Smith said.
At a user panel on cloud migration, representatives from Evernote, Lush and Planet Labs discussed their embrace of higher-level services after the initial lift-and-shift phase, and how they're willing to go deeper on Google Cloud Platform, despite potential limitations to move off that cloud.
"The speed we get in feature development trumps any concern we have with lock-in," said Anirban Kundu, CTO of Evernote, based in Redwood City, Calif., which wants to replace its own database with Cloud Spanner, as well as several other managed analytics services.
The ability to offload those underlying responsibilities is one of the most attractive parts of the cloud, said Troy Toman, director of engineering at Planet Labs. The Earth imagery company was on AWS, but has moved most of its workloads to Google Cloud Platform. Google's cloud also stands out because of its embrace of open source tools.
"I was very attracted to the fact that they support the HBase interface, the fact that you can move a Kubernetes cluster or the fact that TensorFlow isn't limited to Google Cloud," Toman said. "That gave us some confidence that the lock-in was lower."
Price cuts, again
Despite all the attention on higher-level services, there was a nod to the core infrastructure and the area Google first made its play in infrastructure as a service -- lower prices.
A new pricing scheme enables customers to gain considerable discounts if they're willing to make long-term commitments to the platform. Committed Use Discounts reduces listed prices by up to 57% if a customer also buys a certain volume per month for three years. The service is in response to enterprises that still prefer to budget their IT expenses on an annual basis. Users of the service won't be locked into a VM size or type, and there's no upfront cost; customers pay per month for a predetermined aggregate usage.
"We have a big AWS bill," Raby said. "The problem is we'd probably have a hard time moving everything over [to Google Cloud Platform]. But if you could, say, cut it in half, that's a hefty chunk of change."
Overall, Google Compute Engine was cut by up to 8%, depending on the region, while Google added new regions in Canada, California and the Netherlands, bringing the total number to 17 worldwide.
Trevor Jones is a news writer with SearchCloudComputing and SearchAWS. Contact him at [email protected].
Get familiar with Google's big data services
Find the right provider for your big data project
Track the progress of Google Cloud Platform