Since Diane Greene (pictured) took over as senior vice president of Google Cloud in late 2015, we’ve seen a big transformation to make Google Inc.’s cloud more enterprise-friendly. I had a chance to chat with her briefly on Monday, and it’s clear she is proud of Google’s progress.
But the key questions everyone is asking suggest lingering doubts about her chances: Is Google moving fast enough? And can it ever break into the top three cloud suppliers behind Amazon Web Services and Microsoft Azure?
AWS has millions of active customers using its cloud service every month. According to AWS, these millions of customers do not include Amazon.com but are startups, and the largest enterprises, government agencies, and universities. AWS set the bar high with respect to customer testimonials where almost half of the 600-plus sessions at its AWS Re:Invent conference last fall had customer participation.
By contrast, Google’s approach is to “run like Google” at web scale for the enterprise, even though the reality is that most enterprises simply don’t run like Google and probably can’t. Google Cloud consists of only Infrastructure as a Service and Platform as a Service with tools to support both IaaS and PaaS services. Its primary SaaS offering is G-Suite, formerly Google Apps for Work, and it isn’t generally considered part of Google Cloud.
Although Google is not as strong on the breadth and manageability of its platform as Amazon, Google is poised and will be pushing hard to be the cloud of choice for cloud-native applications and big data computing. This will be the starting point in their pursuit to conquer the cloud given the Google history of having a global infrastructure powering Google diverse large-scale products. Unfortunately, many customers we talk to want flexibility in a variety of use cases that relates to the needs of the enterprise.
AWS already has a stronger, enterprise-ready platform and has excellent manageability. To be competitive in the enterprise, Google needs to move faster to fill in their holes in their platform.
At an analyst welcome event In San Francisco Monday, Google put up its obligatory logo slide (pictured) showing some nice adoption. But the question is in what capacity are these customer exclusively using Google versus AWS or Microsoft Azure. To me, the real data Google Cloud can show is competitive wins versus the competition.
Many of the hallway conversations this week in San Francisco indicate that some of Google Cloud’s biggest success stories are also AWS customers. One high-profile example of a competitive win that proved to be less than exclusive was Snap Inc. In Snap’s initial S1 filing, it reported that its operations ran chiefly on Google Cloud, to the tune of $2 billion over five years, but an amended filing revealed a deal with AWS had a billion-dollar, five-year deal as well.
Customer references that are competitive wins vis-a-vis AWS and Microsoft Azure will be the leading indicator of Google’s true traction in cloud share.
The security gap
This will become a very important point as enterprises have unique requirements than typical public cloud or, as the industry calls it, cloud-native. One area of importance for large enterprises is security.
According to our sources in the cloud community, Google has completed a majority of the certifications needed for enterprises but still have another 20 percent to go to be 100 percent compliant. One of the big questions that came up at the Google analyst summit Tuesday was how Google Cloud customers control where their data is stored. This highlights a big security challenge for Google to overcome, as data sovereignty and encryption is something enterprises want to manage and control.
AWS provides data encryption at rest and in transit for its services with key management through AWS Key Management Service or customer hardware security modules. AWS allows customers to use various encryption options for data in Amazon S3 or on-cluster HDFS processed by Amazon EMR. Google’s BigQuery and Dataproc services also have encryption capabilities. (note: previously I stated big query and dataproc didn’t have encryption thx to @thetinot who highlighted that Google has encryption by default).
According to the latest Gartner report, Google’s feature set and scope of services is limited and is missing key capabilities important for established organizations. Specifically, it’s still missing granular and customizable role-based access control and complex network topologies equivalent to those expected in enterprise data centers.
Although Google has strong existing infrastructure capabilities around massive compute instances, AWS provides a much broader selection of compute instances, which means that AWS can provide compute instances that are appropriate for the widest range of workloads. Google workloads have been limited to their internal uses cases and just recently has Google started to sequence to broader, more enterprise-ready workloads. Moreover, Google is amplifying the G Suite as part of the cloud offering, which seems awkward, as most enterprises won’t consume all of the G Suite offerings.
Will AI Change the Cloud Game?
The big drivers for AI are: compute, data and frameworks. The bright spot at Google Next will be the momentum in machine learning fueling a new set of software capabilities that leverage not only data but also data sets. Both AWS and Google have good technology for deep learning but AWS has more experience in production with big data analytics and machine learning. Google is hardcore on deep learning and given its historical affinity with moonshots, I expect it to do well here. The question of whether it will it be a game-changer or just a marketing angle is too early to tell.
Data. Data. Data. And IoT
The market wants technology and software that lets customer data flow to where it needs to be and provides the right tools for the job. AWS has more cloud customers generating most of their data in the cloud today, or are aggressively moving it to the cloud as quickly as possible. The AWS cloud has been supporting Exabyte scale customers who will naturally move AI to that data. Additionally, the AWS cloud can support all AI engines, not just one flavor, which is more appealing because it offers more flexibility and scalability for the enterprise.
AWS has a longstanding core competency in storage and scaling services into big data and analytics. Some of its most successful and fastest-growing areas have been in big data, analytics and Internet of Things. AWS was the first to offer object storage and elastic compute with the launch of Amazon S3 and Amazon EC2 in 2006. Since then, AWS has managed Hadoop, with Amazon Elastic MapReduce, managed petabyte-scale data warehouse, with Redshift in 2012, and provided a fully managed service for real-time processing of streaming data, with Kinesis.
AWS analytics has Athena, Redshift, EMR, Spark, Glue, Elasticsearch and streaming data with Kinesis. For databases, AWS supports lots of engines including open source engines, MySQL and PostgreSQL compatibility on Aurora, plus Oracle and SQL Server, caching and NoSQL with DynamoDB. Google’s sparser offerings are MySQL, plus the recently added Cloud Spanner. AWS has been moving fast to the provide edge computing capability and Internet of Things. Applications that extend to the edge of the network will lead the IoT market.
To give Google credit, it’s making fast progress. The progress Google Cloud has made since Diane Greene came on board is literally night and day from a few years ago.
But as an alternative to AWS, it’s simply light-years behind being ready for primetime with respect to an enterprise-grade cloud. At this stage of the game in terms of compute, storage, networking, security, and manageability, Google is pretty much a one-dimensional cloud player. The only solid use case is a fully managed Google-like environment leveraging Google-only services from IaaS, PaaS and the Google G Suite SaaS.
In my analysis, AWS is so far ahead that Google isn’t even in Amazon’s game. Instead, it’s battling for second, third or fourth place againstIBM Corp., Microsoft Corp., Oracle Corp., and others.
Photo: Robert Hof
Join our 2-days of live coverage with streaming analysis and commentary, on Wednesday and Thursday, March 8-9, at siliconangle.tv