Andy Jassy: 8 AWS Technology Investment Bets For 2020
‘We have a plan that is very ambitious...over the next few years,’ AWS CEO Andy Jassy says. ‘You can expect that you will see a significant amount of continued investment in all of our kind of key building block services. You can expect us to continue to invest very aggressively in compute and storage.’
When it comes to Amazon Web Services technology investments for 2020, CEO Andy Jassy says look at “every area you can imagine.”
The cloud computing provider, known for its fast pace of innovation, has a “very broad group of initiatives rolling right now,” Jassy told CRN in an exclusive October interview in Seattle.
“We have a plan that is very ambitious…over the next few years,” Jassy said. “You can expect that you will see a significant amount of continued investment in all of our kind of key building block services. You can expect us to continue to invest very aggressively in compute and storage.”
Here are some of those areas where AWS plans investments, according to Jassy, who said, “(It’s) a very small portion of what you'll hear from us at re:Invent and over 2020.”
Amazon Elastic Compute Cloud (EC2)
“We're continuing to see customers wanting more and more variants of EC2 instances, and so I think we'll continue to work…to do so,” Jassy said. “They have a lot of additional networking capabilities that they want that we will continue to provide.”
AWS’ core EC2 became generally available in 2008 after launching in beta in 2006. Last month, AWS “chief evangelist” Jeff Barr announced that new AMD-powered and compute-optimized EC2 C5a and C5ad instances built on the AWS Nitro System would be launched soon to run batch processing, distributed analytics, web applications and other compute-intensive workloads. The instances will be available in eight sizes and bare metal form, with up to 192 vCPUs and 384 GiB of memory, he said.
New EC2 instances come down to listening to customers and their requirements for running applications as efficiently – and cheaply -- as possible, according to EC2 vice president David Brown.
“What we're doing there with our...180 different instance types today is trying to meet all those different use cases,” Brown said “Whatever the workload is…we have an instance that can support it.”
On the networking side, with some of its newer features, AWS is addressing how the network can drive innovation and agility within an enterprise.
“Historically, networks have slowed companies down -- engineers go and build something, and the last thing they typically do is think about how do I actually update the firewall and actually deploy this thing and get it into the network,” Brown said. “What we want is a network design where the networking team has the control that they need to provide the security and manage the network, but developers are kind of set free from that to be able to develop their applications and deliver the applications as quickly as they possibly can.”
“On the container side, it's incredible how fast it's growing,” Jassy said. “One of the most exciting things, for customers and for us, too, has just been their adoption of (AWS) Fargate, which is a serverless abstraction that allows customers to be able to run at the task layer, instead of having to worry about servers and clusters underneath the containers.”
AWS last month announced three new features to address customer pain points when building and operating sophisticated container-based applications. Managed node groups for Amazon Elastic Kubernetes Service make it easier to add worker nodes -- EC2 instances -- that provide compute capacity for clusters. Users can create, update, scale or terminate nodes for a cluster with a single command. AWS FireLens is a custom log routing feature that allows users to forward container logs to storage and analytics tools -- without changing deployment scripts, manually installing extra software or writing more code -- by configuring their task definition in Amazon Elastic Container Service (ECS) or AWS Fargate. AWS also added Amazon Elastic Container Registry EventBridge support based on customer requests to be able to start a build process when new container images are pushed to Elastic Container Registry.
AWS also plans on continuing to build its serverless capabilities, according to Jassy.
“On the serverless side, there's several hundred thousands of customers -- active customers -- who are using (AWS) Lambda and our array of serverless capabilities, where it's not just having event-driven serverless compute, but it's having that compute be triggered by all the other services in your platform,” Jassy said. “We have about 60-plus of those services that you can trigger from, which is about three or four times what you can find elsewhere.”
AWS recently debuted AWS Step Functions to allow customers to add serverless workflow automation to their applications. The feature makes application development simpler and more intuitive, because it translates workflows into a state machine diagram that is easy to understand, explain to others and change, according to AWS.
AWS also announced that AWS Lambda -- an event-driven, serverless computing platform -- now supports destinations for asynchronous invocations, a new feature that gives customers visibility to asynchronous invocation results and routes the results to an AWS service without writing code.
And AWS Lambda now supports four failure-handling features for processing Kinesis and DynamoDB streams -- bisect on function error, maximum record age, maximum retry attempts and destination on failure – allowing users to customize responses to data-processing failures and build more resilient stream processing applications, according to AWS.
The database space is not just Aurora, AWS’ fully managed relational database engine that's compatible with MySQL and PostgreSQL, Jassy said.
“As we've been sharing for a little bit of time, most modern technology companies today are not using relational databases for all their workloads like they used to,” Jassy said. “They're really trying to find these purpose-built databases that do whatever the major constraint of their application is (and do) that well, whether it's a...very fast, high-throughput, key-value store or an in-memory cache or a graph capability across different social graphs or having a ledger that's transparent and immutable and cryptographically verifiable or a time series at the edge. All of those are our databases that we have built to allow people to have the right tool for the right job, so I think you'll continue to see us investing there.”
Recognizing customers’ desire for database solutions for specific workloads, AWS last week said it was offering the beta of a purpose-built database specialty certification for builders with expertise in on-premises and AWS databases. It’s the first AWS Certification focusing specifically on database technology. The beta exam can be taken this week at re:Invent and through Jan. 10 at testing centers. The standard version of the test is expected to be available in April.
Analytics And Machine Learning
Analytics and machine learning will continue to be “giant” areas of investment for AWS, according to Jassy.
“We probably have about twice as much machine learning as you'll find anywhere else right now, but it's still such early days in what customers are going to be able to do with machine learning,” Jassy said. “We have a very broad offering that you'll continue to see expand.”
Last week, AWS said that Amazon Aurora is now natively integrated with two AWS machine learning services – Amazon SageMaker, a fully managed service that lets developers quickly build, train and deploy custom machine learning models; and Amazon Comprehend, a natural language processing service that uses machine learning to find insights in text.
And users of Amazon Athena -- an interactive query service that makes it easy to analyze data in Amazon S3 using standard SQL – can now run SQL queries across data stored in relational, non-relational, object and custom data sources. Amazon Athena also now supports user-defined functions, which allow customers to write custom scalar functions and invoke them in SQL queries.
“We're really excited about robotics,” Jassy said. “A lot of tasks that humans have been doing that haven't been very value-added for humans and aren't the best jobs for them are going to be done by robots over time. The number of customers who are innovating and experimenting on top of (AWS) RoboMaker, which is our robotic offering, is quite exciting.”
Released in November 2018, AWS Robomaker provides a robotics development environment for application development, a robotics simulation service to accelerate application testing, and a robotics fleet management service for remote application deployments, updates and management. Last month, it started supporting conditional over-the-air deployment for fleet management, a feature that lets developers specify a pre-check condition during deployment so they have more granular control over the deployment process.
In early 2017, AWS introduced Amazon Connect, its cloud-based contact center service designed to makes it easier for organizations to deliver better customer service at lower costs.
“It's off to an unbelievable start, and customers have so much more they'd like to see us add there,” Jassy said. “It's a very different type of capability, because it's much more cost-effective and performant and capable. Amazon uses it for its own customer service, because it was built from the ground up with cloud and machine learning in mind. They can expect us to continue to add capabilities there, too.”
AWS added web and mobile chat for Amazon Connect last month, which allows organizations’ customers to choose between using chat or making a phone call to get their questions or concerns addressed.
Another new feature, Customer Voice Stream, allows organizations to stream all audio to and from their end-customers in real-time -- audio coming from the customer and a second stream of audio that the customer hears. If media streaming is enabled, the audio is sent to an Amazon Kinesis video stream, and third parties can access it for tasks including real-time text transcription and sentiment analysis for immediate alerts if abusive speech is detected.
‘There's additional things...capabilities that customers would like us to provide on the security side in terms of external security services,” Jassy said. “People are very excited about things like (Amazon) GuardDuty and Macie and Inspector and...would like us to think about additional capabilities in those areas as well. (AWS) Security Hub was a good example of something we just added over the last few months -- things that allow people to have abstractions, so they don't have to do all the stitching together themselves.”
AWS announced the general availability of AWS Security Hub in Boston in June, at its inaugural re:Inforce conference on security, identity and compliance. The new security service gives customers a central place to manage security and compliance across an AWS environment by aggregating, organizing and prioritizing security alerts known as “findings” from AWS services including Amazon GuardDuty, Amazon Inspector and Amazon Macie in addition to AWS Partner Network solutions.