reCap of the AWS reInvent: From the Perspective of a Database Enthusiast

My Personal View on the AWS ReInvent Database Economy

In the following, I will share some impressions of these days based on my personal view with a focus on the latest trend of the databases that can run on AWS, while the native AWS database news is well covered in the AWS reInvent sessions.

The most surprising one, basically every database provider remains the fastest!

Looking at the NoSQL database vendors, with Aerospike, Couchbase, DataStax, Elastic, MongoDB, and Redis the major players were present and had a lot of new features to present. For instance MongoDB the Atlas vector search the Datastax Vector Search the Aerospike Vector support or the Couchbase columnar storage engine for real-time analytics.

New Vector Databases

It becomes obvious that the AI hype has also reached the database world and many providers add vector search capabilities to their database systems. In addition, many dedicated Vector database systems appear on the market, and two of them offered as DBaaS were also present for the first time at reInvent

• zilliz
• and Pinecone.

Both promise to be the optimal solution for use cases with a large scale number of vectors of huge dimensions. It was great to see that zilliz is following a similar approach as we at benchANT do, when it comes to providing a transparent and reproducible performance ranking. Similar to our open database performance ranking, zilliz published a performance ranking for vector databases.

Serverless

Another interesting new database technology that caught my eye is FaunaDB which provides a serverless DBaaS for a distributed document-relational model.

Speaking of serverless, the team of Momento provided a great tool for keeping track of the most precious swag with their swaghunt service. Momento came out of stealth mode a year ago and provides a serverless caching platform (of course with Vector support) with the claim to provide truly elastic serverless and to take on the big players such as Redis or AWS ElastiCache.

Looking at the serverless database trend in general which was a prominent topic at AWS reInvent 2022, this year it is challenged by the vector search support while there are still prominent sessions around the AWS serverless database offers, such as Using Aurora Serverless to simplify manageability and improve costs

Time-Series Databases

It was also interesting to see that from the still evolving time series databases landscape only InfluxDB was present, highlighting their brand new Influx Cloud v3 which is a complete rewrite of the previous InfluxDB V2.

Relational Databases

Apart from the NoSQL databases, there were of course many modern relational database (or NewSQL) providers present, for instance, Cockroach Labs, PingCAP (TiDB), and SingleStore.

CockroachDB promises to be the highly scalable PostgreSQL replacement in a multi-cloud context, PingCAP offers TiDB as a service with the claim to provide a highly scalable or even serverless MySQL alternative for HTAP workloads. SingleStore highlighted its vector search engine in addition to its HTAP features.

DBaaS

And finally, the multi-cloud DBaaS providers Aiven and Instaclustr were presenting their latest service portfolio extensions when it comes to providing open source database (and more) technologies as enterprise-grade cloud offers. And of course, Aiven and Instaclustr also support vector search via the pgvector extension of their PostgreSQL DBaaS offers.

Outro

And finally, the AWS rePlay party has been a remarkable finale of the event, providing fun-filled activities such as a dodgeball arena, the giant Datadog slide, table tennis, pickle ball, and of course great live music acts.

Looking forward to AWS reInvent 2024!

Best Daniel, CTO & Co-Founder