Management Consultant Peter Kiernan joined Amazon Web Services CEO Andy Jassy’s opening keynote speech as one of 400,000+ individuals who had registered to attend the event. The theme this year was reinvention, focusing on the need for businesses to rise to the challenge and continue to reinvent themselves. The keynote centred around products and services that AWS believe will enable their customers to undertake and meet this challenge.
Compute and storage
The staple of the AWS service catalogue has always been the wide selection of compute and storage services that are available. This year brings a new range of Amazon EC2 Instance Types, including macOS images, and the high memory instances for SAP with availability of 24TB. The Graviton2 family of Arm powered EC2 instances will be joined by a new option, the C6gn, which will support enhanced networking. AWS are proud of this family of instances and continue to cite savings of up to 40% when compared with other EC2 instance types.
Machine Learning (ML) emerged as a key area of focus this year, with announcements of ML optimised instances such as Habana Gaudi-based and AWS Trainium showing the commitment that AWS are making to this space.
For compute containers, AWS claims that two thirds of containers that run on the cloud run on AWS. As a precursor for later announcements, Jassy revealed two new services that will allow customers to run Amazon Elastic Container Service (ECS) and Amazon Elastic Kubernetes Service (EKS) containers within their own on-premises data centre.
He also announced the introduction of a new Amazon Elastic Block Store (EBS) volume type called gp3. This is a Solid State Drive (SSD)-based offering which promises a cost reduction of 20% per GB and a 4x peak uplift in throughput. A Storage Area Network (SAN) service built for the cloud, io2 Block Express is an interesting proposition, advertising 256,000 input/output operations per second (IOPS) compared to a General Purpose SSD, which only offers 16,000 IOPS per second.
Amazon claim to have migrated 350,000 databases to AWS. We are particularly excited to test their new service, Babelfish for Amazon Aurora PostgreSQL, which promises to expedite the migration of Microsoft SQL Server databases. Finally, AWS Glue Elastic Views will facilitate the synchronisation and querying of data sets from the plethora of different datastores that Amazon currently offer without the need for writing custom code.
Machine Learning (ML)
As noted earlier in this blog, machine learning was presented as a key topic during the keynote speech. Jassy highlighted that within Amazon SageMaker alone, over 50 new capabilities have been added over the past year. Several new services were announced that would continue to evolve these technologies, including:
- Amazon SageMaker Data Wrangler, which promises to recommend and execute transformations based upon the selected dataset.
- Amazon SageMaker Feature Store, which will attempt to make it easier to name, organise, find, and share features. There is a strong focus on making ML more accessible.
- Amazon SageMaker Pipelines, which is a CI/CD pipeline for ML.
- Amazon Connect, the cloud-based customer contact centre. The impact of Covid-19 has seen an additional 5,000 deployments of this service alone. The aim is to use Amazon’s ML technologies to enhance the service in a variety of ways.
- Amazon Connect Wisdom, which promises to provide agents with product and service information to solve issues in real-time;
- Amazon Connect Customer Profiles, a service that helps surface relevant customer information to a call handle;
- Real-Time Contact Lens for Amazon Connect, which identifies issues in real-time for customer interactions. It promises to identify when a call is going wrong and offers supervisors an opportunity to provide coaching or step in to resolve the issue
Redefining hybrid cloud
Amazon shared interesting views on the nature of hybrid cloud in 2020. Building upon their VMware Cloud on AWS and AWS Outposts services, announcements were made for new, smaller variants of AWS Outposts. It is considered that these smaller devices could be deployed in retail locations or hospitals where strong network connectivity may present a barrier to entry and an entire AWS Outpost stack would be excessive. An additional service that is moving the AWS stack closer to the customer is AWS Wavelength, which supports ultra-low latency applications for 5G networks.
These announcements, along with the ECS and EKS Anywhere services, represent an interesting change for AWS and we are interested to learn the possibilities that these services will open up for customers.
In a nutshell
Drawing the keynote to a close, Andy Jassy left the audience with the key message that successful customers must build their capacity to make changes, concluding that invention and reinvention are both possible.
This year’s AWS keynote detailed a range of services that will aid customers on their respective journeys to reinvention. With an additional four keynotes to go, we are excited to learn more on AWS service offerings over the next coming weeks and the benefits that these services will present for customers.
If you would like to discuss any the above topics in more detail, please get in touch with one of our AWS experts.