LAS VEGAS -- A slew of AWS enhancements unveiled by Amazon this week will likely resonate with enterprise customers who have invoked various excuses to avoid the platform -- but even if they have a hard time still saying no, many will need guidance to decide what's best for them.
Offerings unveiled here at AWS re:Invent -- new instance types, PostgreSQL support in Amazon Relational Database Service, a further push into hybrid cloud and more machine learning options -- represent efforts to lower the bar for enterprise adoption and also catch up to emerging competition.
Variety breeds opportunity and complexity
Public cloud detractors often cite limited instance types as a primary reason not to move to the cloud, but Amazon has countered by gradually adding more options for VMs. Amazon upped the number of instance families to 11, added a virtual private server option, expanded options within T2 and R4 instances, and said it will incorporate the next-generation Intel processor in C5 instances.
The new F1 instances offer customer-programmable field-programmable gate arrays, while the I3 is targeted at I/O-intensive relational databases. Customers can also use Elastic GPUs to attach the graphic accelerators to an existing Elastic Compute Cloud instance.
AWS enhancements are generally a good thing, but the slew of options raises concerns of customer confusion. This could be addressed by system integrators, but it could present extra strain as less savvy enterprises start to move to AWS.
"The highly technical companies will love it and figure it out, but someone who's just looking to get some work done may have trouble figuring out which one to choose," said Mike Kavis, vice president and principal cloud architect at Cloud Technology Partners, an AWS consulting partner in Boston.
Within the next year, he said he hopes to see a layer of abstraction that can help customers select the right instance for them.
One advantage for Amazon, however, is customers in the cloud can test out the various instance types to figure out which is best for them, said Rich Sutton, vice president of engineering at Proofpoint Inc., a security and compliance provider in Sunnyvale, Calif., and an AWS customer.
"I don't perceive there to be too many instance types," Sutton said. "They all have their tradeoffs, but unlike the old hardware days where you had to lay out Capex to figure out which is the best for you, you can just try stuff out."
Hybrid cloud seeps into AWS strategy
Amazon also made its biggest commitment yet to hybrid cloud. AWS Greengrass, unveiled at AWS re:Invent, embeds AWS Lambda compute and services into connected devices that can be run locally. While the latest version of the popular Snowball data transfer service, Snowball Edge, can store 100 TB of data and includes compute and storage capabilities inside the device, Snowball Edge includes Greengrass, too.
Snowball Edge is a server-like device, and while Greengrass is currently focused on internet of things, this combination can be expanded to more general uses as customer demand increases, said John Rymer, vice president and principal analyst at Forrester Research.
"It's not that Amazon is going into the private cloud business, but it's the most aggressive that they've ever been in inserting their technology into [customers'] data centers," Rymer said.
Amazon also literally rolled out another new data transfer service, Snowmobile -- a 45-foot-long semi-truck, which was driven onto the conference hall floor to close out its Wednesday keynote. The service will allow customers to transfer exabytes of data to AWS with a 10 Gbps line.
Amazon estimates it can speed up a transfer that would take more than a quarter century down to six months. It's unclear how many customers actually have enough on-premises data stores to require such a massive device, but the consensus among show attendees speaking with SearchAWS is this likely will be a niche service, at least in the short term.
AWS enhancements catch up to the competition
Amazon continues to dominate the public cloud market with a platform viewed as more mature and with a wider breadth of features than any of its competitors. While some of the new AWS enhancements continue to extend the company's innovation lead over rivals, others attempt to play catch-up.
Amazon Athena will allow customers to perform ad hoc SQL queries in Simple Storage Service without the need to spin up or manage any additional infrastructure. The service can provide results in milliseconds and works for either batch or real-time analytics. In many ways, Athena is Amazon's answer to Google BigQuery, which has been one of the biggest drivers of Google Cloud Platform adoption.
Mike Kavisvice president and principal cloud architect, Cloud Technology Partners
There are also improvements to artificial intelligence capabilities on AWS, including machine learning services for speech and facial recognition. Those capabilities already can be found on IBM Watson, Google Cloud Platform and Microsoft Azure.
"When they were describing the machine learning updates, Amazon kept hitting on quality, quality, quality," Rymer said." That's the kind of thing you say when you're behind and you've got to have an edge."
Overall, the feature disclosures were seen as positives and help further AWS' strategy to remove reasons not use the platform, even if that means matching what other providers offer.
"Even though they have more feature function than anyone, those areas where they do have gaps they're starting to address," Kavis said. "They're starting to remove the excuses for you to go other places."
Trevor Jones is a news writer with SearchCloudComputing and SearchAWS. Contact him at [email protected]
Your comprehensive guide to AWS re:Invent 2016
AWS expands enterprise migration capabilities
Using AWS in a hybrid cloud makes sense to customers