As an AWS Premier Consulting Partner, KCOM has been attending to AWS re:Invent for a number of years. Each year, AWS announce and unveil a swathe of new technologies within their service offering, often pushing the boundaries of expectation, but also filling in gaps in their portfolio.
This year, amongst the KCOM attendees were Chris Cox, Product Innovation Lead and myself, Wayne Stallwood, AWS Practice Lead. These are our thoughts on the top five announcements they have seen in the sessions and keynotes. We have also included a viewpoint from Nick Holmes, Consultant within the Cloud Sales team.
Highlights from Chris Cox
With the announcement earlier this year of Amazon Connect, AWS looked set to disrupt the contact centre market. The addition of Amazon Transcribe, a voice to text service that will soon be able to recognise different actors within an audio stream, shows their commitment to this space and reinforces how they are set to disrupt this sector.
Machine Learning is immensely powerful for allowing companies to supercharge the value of their data, but it’s also extremely complicated. Not so much anymore. SageMaker takes huge amounts of the heavy lifting out of a Machine Learning setup, which will enable companies to vastly lower their cost of entry.
While it wasn't the most ground-breaking announcement, AWS Cloud9 has the potential to be a real money saver. Increasing the efficiency of developers and having fully integrated access to the entire AWS stack from the IDE should really increase the output of engineering teams.
IoT Device Defender
Security for IoT has been a hot topic since it was discovered that millions of IoT devices around the globe had been hijacked into a botnet that was causing denial of service attacks on huge portions of the internet. Natively available IoT security with simple, scalable provisioning makes enterprise IoT on AWS an even more attractive proposition.
A graph database was really the only major database technology AWS was missing. Well, not anymore!
Highlights from Nick Holmes
I was really interested to hear that Apache ActiveMQ was being productised into the Amazon service catalogue. I have been using SNS and SQS for over four years now, but was acutely aware of the missing ‘standard’ messaging protocol support in the turnkey offerings.
This will make a big difference with migration projects as it provides the opportunity to either avoid using non cloud-native queues completely, or gives a target to move to as the ‘next step’ once the migration is established. The choice remains with the customer, but it really improves the foundations provided by Amazon for wholesale migration into the cloud.
Elastic Container Service for Kubernetes
With containerisation front-and-centre for anyone looking at moving to a ‘Serverless’ platform, but unwilling or unable to take a bet on moving to AWS Lambda, or equivalents in other cloud platforms, the announcement of the Elastic Container Service for Kubernetes (or K8s) makes the Kubernetes containerisation platform a first class citizen in the Amazon offering. This has major implications, as workloads are suddenly compatible, at the container level, between the various clouds, and within AWS can be integrated in all the ways you might expect with AWS IAM, VPC and other services.
With the exponential growth in stored data, putting that information to work is increasingly important. SageMaker brings two features to the table that we find equally exciting: firstly, the ability to work with ML engines in a turnkey environment and make sense of owned data, and secondly, making the use of ML sufficiently straightforward that owned datasets gain more value as customers are able to market these to newly tooled-up consumers of their data.
Amazon API Gateway Private VPC Endpoints
As a proponent of the Serverless approach, and a user of API Gateway, I was really pleased to see this small change, as I think it will have far reaching impact, allowing applications to be further protected from external sources, while making the best use of Serverless or existing applications. Probably the most compelling feature is the ability to use PrivateLink to connect back to other internal resources behind the API Gateway.
The addition of local resource access and ML Inference integration to Greengrass goes a long way to making Lambda Beyond the Edge an even more useful service. The ability to remotely update the core, as well as the functions deployed on the Greengrass node, provides a standardised means of updating in the field, allowing for far more secure IoT deployments, with field upgrade capabilities considerably enhanced.
My own key takeaways
Aurora Multi-Master brings a level of availability to MySQL and Postgres Databases that even the most established commercial database offerings struggle to provide. Near instantaneous cross Availability Zone failover is a really big deal.
Aurora Serverless, the world’s first truly autoscaling RDBMS as a service! Capacity is billed by the second and scale up events can happen in five seconds. Being able to scale back to zero (paying just for storage) opens opportunities to cost effectively manage a whole new type of workload.
Replicating data across AWS Regions has always required specific intervention or tools. DynamoDB global tables is a very welcome addition, which greatly simplifies our multi-region architecture.
Amazon GuardDuty is a fantastic addition to the ever expanding arsenal of security features provided within Amazon's cloud platform. Now all customers, large and small, can apply the new generation of intelligent threat detection and of anomaly detection driven by machine-learning.
Managed Kubernetes (EKS) and AWS Fargate working together will significantly reduce the complexity of deploying and effectively scaling container based solutions. We have been waiting for this and it has an immediate impact on projects we are working on today.
If you're interested in all things AWS, or would like to discuss any of the new technologies, please feel free to drop us a line using the form below - we'd love to hear your thoughts too!