Experts expect AWS re:Invent 2019 to spotlight AI, security and more
Re:Invent 2019 will yet again be heavy on product rollouts. Get ready for the show with a rundown on the topics and technologies our experts predict AWS will spotlight.
As the AWS re:Invent 2019 conference approaches, the cloud computing community can expect another round of product rollouts and feature updates as AWS sets the agenda for another year.
Last year's re:Invent included some predictable product rollouts, such as Amazon Managed Blockchain and AWS Transit Gateway, as well as the typical expansions to EC2 instance types. But AWS showed it can also throw a curveball that even the best pundits can't see coming when it disclosed plans for AWS Ground Stations.
So what should users and attendees expect from re:Invent 2019? Our SearchAWS contributors expect developments regarding machine learning and AI, along with multi-cloud and security enhancements. See the rest of their predictions below.
Kurt Marko, consultant
Last year, AWS emphasized improvements for enterprise adoption by embracing hybrid infrastructure with VMware and Outposts, new features for VMware on AWS and extra-large instances for workloads like SAP. I expect to see more of the same this year. AWS always reveals several surprising breakthroughs, which given its enterprise focus, will come in the form of one or more managed services for database applications, such as ERP potentially partnering with SAP or CRM to compete with Salesforce and Dynamics 365.
Outside of improvements for core enterprise workloads, I expect continued innovation in AWS' AI and machine learning portfolio with new services in both categories. This might include more convenient features for AI frameworks, such as TensorFlow and MXNet. Look for improvements to AWS Deep Learning Containers, perhaps with integrations for Fargate and Lambda. These integrations would provide a streamlined process to deploy and integrate containerized apps with other services.
In addition, AWS has an impressive capability for developing custom hardware, with offerings such as the Gravitron processor, AWS Inferentia and Nitro system. While AWS is not likely to release an entirely new processor architecture this year, I expect significant improvements to both the ARM and machine learning processors, with stronger capabilities that will process x86 instances for higher-end workloads.
David Linthicum, cloud analyst
I think that AWS will surprise the cloud computing world with support for multiple public clouds, or multi-cloud, as part of their new technology announced at AWS re:Invent 2019. Google has embraced the use of multi-cloud with the launch of Anthos earlier this year, as customers are more likely to leverage heterogeneous public clouds. I expect AWS to embrace this trend as well, understanding that most AWS customers need built-in support for multi-cloud, and are likely to select cloud providers that support other public clouds as well.
George Lawton, IT writer
In 2019, AWS will double down on security by default. In recent years, the vast majority of cloud breaches have involved misconfigurations of AWS tools and features. Politicians, including Sen. Ron Wyden, have even started inquiring into whether AWS has done enough to secure its cloud platform.
Although AWS is not likely to address these concerns directly, I expect the cloud provider will announce a number of new security enhancements to its service for enterprises, such as security by default and better security scanning capabilities.
Brian Kirsch, IT architect
I think this year should be interesting given the recent news of Microsoft winning one of the largest government contracts ever for cloud services. I expect to see this bring out a bit more of the Amazon marketing machine to promote that AWS is still the largest cloud vendor and showcase the extensive features and services the platform offers.
Machine learning and AI will likely dominate the sessions catalog, show floors and conversations. While the number of people that can afford to use many of the machine learning and AI services -- or even have a use for them -- is still small, I anticipate a lot of focus on it.