This year, AWS re:Invent expanded to three weeks in order to better suit the remote work reality many of us are in, and with over 80 significant new features or services, it feels as though the number of announcements scaled to match. There are far too many to do anything but scratch the surface, so while these ones in particular caught our eye, it would be well worth taking some time over the holidays to dive into the others, I know we will!
Some of the highlights that caught our attention this month include Babelfish for Aurora PostgreSQL, AWS Fault Injection Simulator, AWS Managed Services for Prometheus and Grafana, smaller form factors for AWS Outposts, and strong read-after-write consistency for AWS S3. These selected announcements are for services and features that can assist enterprise thought leaders who are looking to leverage the cloud and drive efficiency within their organizations.
Babelfish for Aurora PostgreSQL
License costs are a large part of many environments, and databases can be a significant portion of that cost. As many organizations look to move towards open-source options to combat those costs, they can find themselves stalled by the level of effort required to update their applications and processes to use a new database engine. With that in mind, AWS has released Babelfish for Aurora PostgreSQL, which allows Microsoft SQL Server commands to be run against a database that has been migrated to Aurora PostgreSQL.
With Babelfish, a separate Aurora endpoint is enabled which acts as a translation layer for T-SQL language and semantics. This enables applications to continue working using their existing queries and immediately removes the need for those MS SQL Server licenses, ultimately reducing costs for the organization. The Babelfish endpoint is free, and exists in parallel with the regular Aurora PostgreSQL one, allowing for new or updated applications to begin fully utilizing the PostgreSQL engine.
This release, currently in preview, promises to allow for huge cost-savings opportunities and reduce the effort for organizations making the move away from licensed databases.
AWS Fault Injection Simulator
Many of us do our best to follow best practices as we build our applications and cloud environments to be fault-tolerant and resilient – able to handle an EC2 instance crash, an Availability Zone failure, or maybe even an entire Region disappearing. Testing this well-planned architecture has always been difficult in practice, as simulating some of these potential failures within our AWS environments can be extremely challenging. This month, AWS announced the AWS Fault Injection Simulator (FIS) to help. This service, due for release in 2021, allows for organizations to run controlled experiments against their environments and to gain insight into the outcomes.
Using AWS FIS, experiments are defined and created through templates, which this service uses to simulate increased CPU, API throttling, EC2 instance shutdowns, geographic failures, and more, while maintaining careful control over the experiment as it progresses. The experiment can be run against live environments, from staging to QA to production, with automated guardrails to end the experiment early if it breaches those limits. It’s also possible to add AWS FIS into a CD pipeline, allowing for even greater confidence that your infrastructure meets your requirements on each code release and that you’re ready for disaster should it strike.
Amazon Managed Services for Prometheus and Grafana
Observability was a big component of Dr. Werner Vogel’s keynote presentation during this re:Invent conference, stressing the need to measure and understand the health of our systems, and the need to be capable of collecting these metrics at the vast scales at which we’re now regularly deploying our applications.
Prometheus is a popular open-source monitoring project which is part of the Cloud Native Computing Foundation. Prometheus is capable of high-scale, low latency metrics collection, monitoring, and alerting, and is seeing rapid adoption by organizations of all sizes as they look to gain improved awareness of their applications and services.
Grafana composes observability dashboards powered by various data sources, with Prometheus and CloudWatch being two great examples. These powerful dashboards can be critical in observing the health of your systems, collating multiple dimensions of information, and identifying emerging patterns within your metrics.
The Amazon Managed Service for Prometheus (AMP) enables Prometheus to be deployed in an easily scalable and managed form, allowing organizations to focus on the instrumentation of their code and systems, while AWS handles the underlying infrastructure. Similarly, the Amazon Managed Service for Grafana provides a scalable implementation of Grafana, freeing up your time to focus on using it to understand and correlate the metrics you’re gathering.
AWS Outposts 1U and 2U form factors
As many organizations migrate or build all-new in the cloud, geographic proximity or data residency can be a big limiting factor, restricting their choices or perhaps even preventing their adoption of the cloud in some circumstances. AWS has recognized that many companies want to continue to design and build their workloads in the cloud-native manner they’re using elsewhere. They had previously released AWS Outposts, a custom 42U cabinet to be deployed in data centers, as a solution for this. AWS Outposts allows for treating compute and storage resources in the cabinet as an extension of an organization’s AWS environment, using native AWS APIs and services. For many purposes though, the form factor of such a cabinet was beyond their needs. This re:Invent, AWS announced 1U and 2U form factors of AWS Outposts, which are rack-mountable devices that will allow greater flexibility on where these can be installed and used. These devices, which will be available in 2021, could easily power solutions in retail stores, factories, and shop floors, or hospitals.
Amazon S3 Update – Strong Read-After-Write Consistency
As the oldest production service within AWS, Amazon S3 is a much loved and much-utilized service, important both to other AWS services and to many of the AWS deployments of organizations around the world. S3 powers a wide variety of roles, from serving website and application content, to storing backups and snapshots, to powering big data projects.
Data stored in S3 has always been subject to eventual consistency, essentially meaning that there exists a small period of time between when an object is put into S3, and when it is available to be read from S3, which is challenging for data processes that require the ability to read the data immediately after it is written, big data workloads being a key example.
As of this month, AWS S3 now has strong consistency for GET, PUT, and LIST operations, with no extra cost, and even better, no extra configuration required to enable it. This capability is now available in all AWS regions, has no impact on performance, and makes S3 even more of an essential offering in AWS.
To follow these monthly updates and gain insights on how they can impact your business, subscribe to our blog!