Like the rest of Google’s business-centric services, Google App Engine didn’t start out as a business-centric service. When it first launched in the spring of 2008, it was a rather intriguing way for hardcore coders to run their own applications atop Google’s top secret online infrastructure, and many questioned whether the mainstream developer — let alone big businesses — would ever play by the strict Google rulebook that App Engine required.
Those doubts still linger, but Google is intent on making them go away. On Tuesday, the company announced that it’s now offering “enterprise level” service and support for the three-year-old service. “When choosing a platform for your most critical business applications or standardizing on one across your organization, we recognize that uptime guarantees, easy management and support are just as important as product features,” Google said in a post to its official enterprise blog.
As it works to bring the service to big business — and better compete with the likes of Amazon — the Mountain View web giant has also spruced up a sister service dedicated to storing large datasets, and it’s slowly removing many of the Googly restrictions that threatened to limit App Engine’s appeal. Just last week, the company added an old school relational database to the service, providing an alternative to the NoSQL data model that underpins the Google infrastructure.
Like Microsoft Azure and VMware’s Cloud Foundry, Google App Engine is an online service for hosting applications and automatically “scaling” them to an ever larger number of servers. Unlike an “infrastructure cloud” a la Amazon’s EC2, these “platform clouds” are designed to hide all underlying infrastructure from the user, including virtual machines. Developers code to the cloud’s APIs (application programming interfaces), and the cloud takes care of the rest.
These services promise extreme ease of use, but with App Engine, not everything was easy. In order to provide that automatic scaling, Google tightly restricted how developers could build their applications. It limited the software languages, libraries, and frameworks the service would run, and it insisted on a NoSQL data model. It even limited how long each application process could run.
But this spring, at its annual developer conference in San Francisco, Google announced that App Engine would soon exit its three-year beta phase, and the company began removing some of its chief restrictions. Most notably, the company introduced a beta API for longer-running processes, and now, there’s a relational database beta that companies and developers can use in lieu of its new-age NoSQL data store.
NoSQL databases — including the proprietary BigTable database used by Google — are designed to scale across thousands of servers as a way of handling much larger datasets. But relational databases — which order data into neat rows and columns — give you more ways of slicing and dicing your data.
On Tuesday, Google also removed the beta tag from its standalone storage service — a service that operates alongside App Engine and does little more than store vast amounts of unstructured data. It’s more of an infrastructure cloud, providing a raw computing resource.
The service was originally known as Google Storage for Developers, but it has now been rechristened Google Cloud Storage. Once again, the company is moving away from Joe Developer, with an eye on Joe Enterprise.
The service is similar to Amazon’s Simple Storage Service (S3). “Google Storage is designed for very high availability, performance and scale, and uses the same technology that powers some of the highest traffic sites on the Internet,” a Google spokesperson said, presumably referring to NoSQL. But the company also boasts that Google Cloud Storage includes certain tools you won’t get from the competition, including the ability to readily share data with Google accounts via the Oauth authentication standard.
DNAnexus — a Mountain View-based startup offering a software platform for managing and analyzing human genome data — is now using Google Cloud Storage to serve up what it calls the most comprehensive archive of publicly available DNA data. In February, the National Center for Biotechnology Information (NCBI) — the primary host for public genomic data in the U.S. — announced that it would slowly phase out its DNA archive due to federal budget cuts, and DNAnexus has picked up the slack.
“The government just can’t keep up with how quickly these datasets are growing and how much it costs them to host,” DNAnexus CEO and founder Andreas Sundquist told Wired. “It’s huge database today, but in light of what we’re going to be doing in the future, it’s tiny.” A Google spokesperson tells us that at about 100 terabytes, the DNAnexus “Short/Sequence Read Archive” (SRA) is one of the largest datasets deposited in Google Cloud Storage by any outside organization.
DNANexus is funded in part by Google Ventures — the Mountain View giant’s venture capital arm — and though Google says this played no role in the startup’s decision to use Google Cloud Storage, it does say that worked with DNAnexus on its genome database “to make sure they were successful in their mission.” Google does not officially offer “enterprise” support for its new Cloud Storage service, but one gets the feeling this is on the way.
With App Engine, if you pony up for a $500 a month “Premium” account, Google provides technical support while guaranteeing your application will be available 99.95 percent of the time. Plus, you’re free to run as many applications as you like atop the service. Ordinarily, developers must pay a minimum fee for each new app.