Connecting the dots: Reflections from AWS re:Invent

Two major trends stand out.

  • Tor Lekven
Jan. 2 20233 min. read time
  • AWS

Last years’ re:Invent is now some weeks old, and I wanted to share some thoughts from the event. As before, many new features and capabilities were announced, and looking back I see two major trends that stands out.

The first is that AWS keep moving up the stairs from their beginning as an infrastructure as a service company, to a software as a service company. Although there is still a lot of innovation happening in their infrastructure services, like new EC2 instance types, Graviton processors and VPC lattice, it is also evident that Amazon continues to invest heavily in data governance, AI, analytics and machine learning.


AWS is now connecting their capabilities

My second observation is that AWS is really starting to integrate their services. While new features in previous years have often been more or less stand-alone, AWS is now connecting their capabilities. This is exciting, because the value of these services combined, is very often greater than the sum of each. The examples from re:Invent were too many to mention all, so I have picked out a few that got my attention:

Introducing Amazon DataZone

This year, AWS introduced Amazon DataZone; a data management service that lets you share, search and discover data. It’s integrated with Redshift, Athena and Quicksight, which enables users to access these services seamlessly, as well as common views and data management across the organization, irrespective of where the data is stored.

It also uses machine learning to collect and suggest metadata. Another noteworthy feature is the Amazon Aurora zero-ETL integration with Amazon Redshift, which enables near real-time analytics on transactional data, and the possibility to consolidate data from multiple Aurora databases, all done serverless.

A range of updates to Sagemaker

There has been a lot of buzz around AI and machine learning over the past few years. But adoption so far has been somewhat slow, and driven by a handful of leading-edge companies with specialized developers.

If the technology is to achieve widespread use, it needs to be simplified and less costly, and AWS is doing their bit to democratize it. They announced a range of updates to Sagemaker, and more of their AI and ML capabilities are now available as serverless, which makes it possible to lower the cost of this data intensive domain.

Further, AI capabilities have been incorporated into devops tools like CodeWhisperer, CodeCatalyst and CodeGuru, and Amazon Guard Duty now uses machine learning to more accurately detect suspicious activity. Amazons contact center solution, AWS Connect, is another area where AWS has integrated their AI, now boasting features such as machine learning based forecasting and training tools.

Integrating and improving existing services

In his keynote, Swami Sivasubram, AWS VP of databases, analytics and machine learning, was reflecting on how great inventions happen. In popular belief, they are often portrayed as a “glimpse of light” (like Newton and the apple), but research shows they are usually the result of people connecting the dots after years of accumulating skills, knowledge and experience. It seems to me AWS is adopting the same principle, driving much of their innovations by integrating and improving existing services.

Only time will tell what’s next, but my bet is that these trends will continue. For DNB, the big question is how we can best utilize the range of new services coming from AWS.