Somehow, no matter how fast it grows and how big it gets, cloud computing just seems to keep accelerating. At least, that’s what it looked like at the fifth annual AWS re:Invent conference this week, where some 32,000 cloud enthusiasts have thronged to Las Vegas to learn about what’s new from Amazon Web Services and its ecosystem of partners and customers. (Another 50,000 watched the keynotes via live-streaming.) The whole event has the palpable buzz of a celebration of a successful, growing industry that feels like it’s on the right side of history.
Of course, New Relic is there, as a Diamond Sponsor with a packed booth hosting lightning talks from New Relic experts and partners plus a couple of sold-out sessions with our cloud architect Lee Atchison (who spent seven years at Amazon and AWS)—not to mention a filled-to-capacity after-hours event.
We’ll have more on New Relic’s participation in upcoming posts, but first let’s take a look at some of the most interesting stuff that AWS CEO Andy Jassy revealed in his opening keynote Wednesday morning.
The cloud grows “superpowers”
As is customary for re:Invent, Jassy started off with a list of AWS’s growth statistics, which somehow keep getting more impressive every year. Jassy said AWS now has a $13 billion revenue run rate, representing a 55% year-over-year growth. That comes from “millions of active customers,” he said, ranging from tech startups to giant enterprises and including “every imaginable” vertical market segment, including 23,000 government entities around the world, 7,000 academic institutions, and 20,000 non-profits. The AWS Marketplace, meanwhile, now includes more than 3,500 from 1,100 sellers, and usage is growing 112% year over year.
AWS is the fastest growing $1 billion-plus enterprise tech company, he said, growing 55% year over year. He cited Gartner’s claim that AWS is several times the size of the next 14 competitors combined. And he quoted Fortune remarking on the enterprise’s “almost insatiable demand” for cloud services.
To satisfy that demand, Jassy announced a long list of new AWS features and functionalities for everything from high-performance computing to entry-level applications, and involving tiny IoT devices as well as 45-foot-long shipping containers.
T2, R4, C5, I3, F1, and Elastic GPU
AWS announced a pair of new T2 instance sizes: t2.xlarge and t2.2xlarge has approximately twice the power of the existing Large T2 instance, while 2XL has approximately twice the power of the XL instance.
The company also announced it was upgrading its R3 in-memory processing instance from R3 to R4, once again approximately doubling the compute power and speed. Similarly, I2 is being upgraded to I3, improving IOPS by nine times, and approximately doubling stats such as memory, storage, and CPUs. And to keep the theme going, the 4 instances are being upgraded to C5 using Intel Skylake chips to approximately double performance.
(For more details, see Amazon’s blog post: EC2 Instance Type Update – T2, R4, F1, Elastic GPUs, I3, C5.)
GPUs also got attention, with previews of the new Elastic GPU for Amazon EC2. It’s designed to let users who may need only a small amount of GPU compute buy just what they want and attach it to any AWS instance. Finally, Jassy announced a developer preview of the new F1 instance type featuring field-programmable gate arrays (FPGAs), which can be programmed in the field to boost performance for particular tasks. “All your workloads are not vanilla,” he said, and having the right tools is essential to maximize performance.
How low can Amazon go?
If FPGA represents the high end, the new Amazon Lightsail is a push deeper into the low end of the company’s customer base. Lightsail is designed to let users create simple workloads for simple applications without having to deal with AWS’ inner workings. Developers can access their virtual private servers (VPS) starting at just $5 per month, and grow from there as needed.
Analytics and AI
That focus on simplicity was also apparent in Amazon Athena, a new way to perform simple SQL queries of your data in Amazon S3. The idea is that “everyday developers” can get more utility out of their data, Jassy said, without having to deal with any infrastructure. He said results come in seconds— “sometimes milliseconds”—and users pay only for the queries they actually run.
And it seems like you can’t have a tech conference in 2016 without Artificial Intelligence making an appearance, and re:Invent did not disappoint. Jassy announced three new AI projects, and promised more for 2017.
Amazon Rekognition is an image detection and recognition service powered by deep learning. Jassy said it can tell the number of people in an image (to make sure they don’t get automatically cropped out, for example), perform face matching (useful in security applications) and also detect their sentiment (smiling or frowning). Even better, he said, Rekognition fine-tunes itself the more it’s used.
Amazon Polly is a text-to-speech service designed to take a stream of text and spit out a stream of natural-sounding speech, including correctly pronounced colloquialisms, such as abbreviations and other natural-language assumptions. “The temp in WA is 75 F” might be rendered as “The temperature in Washington is 75 degrees Fahrenheit,” for example.
Amazon Lex uses Amazon’s Alexa technology to help developers build conversational voice and text Interfaces that leverage Lambda triggers to pull in contextual information.
The grass is green in IoT
If there’s any tech trend as hot as cloud computing, it has to be the Internet of Things. And Jassy noted that the two are intimately connected, because many IoT devices have very little onboard capability and rely on the cloud for all kinds of assistance. That’s where AWS Greengrass comes in, designed to bring the power of AWS to simple IoT devices. The Greengrass software can be embedded or downloaded into IoT devices, and use Lambda to trigger events even when not connected to the cloud—which can often happen in far-flung IoT implementations. Applications include smart homes, agriculture, and manufacturing, Jassy said.
The really BIG news
At AWS re:Invent 2015, I was impressed with AWS Snowball, a suitcase-sized data transfer appliance. Apparently, I wasn’t the only one. Jassy said that demand for Snowball was 10 times higher than expected—some customers are even snapping selfies of themselves with the device. The device’s surprising popularity led AWS to develop the Snowball Edge, which doubles capacity to 100TB and adds clustering capability and AWS Lambda-powered local processing to enable new applications; for example, collecting oceanographic or aeronautic data in the field, doing rough analysis on site, and then sending the device back to AWS to put the data in the cloud to allow deeper analysis.
But the keynote audience gasped when Jassy rolled out AWS Snowmobile, a (literally) truck-sized version of the device designed for “exabyte-scale” data storage situations. As a full-sized semi-trailer drove on onstage, Jassy explained that Snowmobile was a “new definition of container computing,” and said it’s designed to be dropped off at the customer’s data center, filled with up to 100 Petabytes of data, and then driven to an AWS data center for transfer into the cloud. Now that’s scale!
The next 10 years
There was much more in the two-and-a-half-hour keynote, including Amazon Aurora database compatibility with PostgreSQL, not to mention customer testimonials from McDonalds, Enel, and FINRA, among others. But the overarching themes centered around continual expansion of the cloud’s capabilities, adoption, and use cases.
“I believe the next 10 years will have markedly more innovation,” Jassy said. During AWS’ first decade of life, a lot of time and angst was wasted on worrying whether people would really use the cloud, he explained. Now, he continued, “there are no more ‘if’ conversations. There are only ‘when’ and ‘how’ conversations.” That will give companies more time and energy to innovate in the cloud, he predicted.
Watch the video of Andy Jassy’s full keynote below: