Imagine stepping into a world where the boundaries of technology are
continuously pushed, a place where innovators, dreamers, and engineering experts
come together to share their latest breakthroughs and accomplishments. This was
the atmosphere at AWS re:Invent 2024, held in the vibrant city of Las Vegas. For
those who might not have had the chance to attend, let me take you on a journey
through our experience.

As we arrived at the event, the excitement was tangible. AWS re:Invent is not
just a conference; it's a celebration of all things cloud, with a particular
emphasis on Generative AI this year (again). Everywhere you looked, there were
engaging sessions, hands-on labs, and booths showcasing the art-of-possible with
cloud. For the Rearc team, this was an opportunity to dive deep into the latest
advancements in the AWS ecosystem, connect with like-minded professionals, and
gather insights that would fuel our passion for innovation and solving deep
engineering problems.

One of the highlights of our trip was the Rearc happy hour on Monday evening. We
transformed a cozy venue into a haven for tech enthusiasts, where our customers,
prospects, and partners could relax, network, and share their stories. The
energy in the room was infectious. As people mingled and exchanged ideas, it
became clear that this was more than just a networking event—it was a community
coming together. With every conversation, we celebrated not only our shared
achievements but also the endless possibilities that lie ahead. The turnout was
incredible, and the feedback we received was a testament to the value of our
collective efforts in driving technological progress into 2025.
Top 10 AWS Announcements at re:Invent 2024
As a tech enthusiast with a background in solution architecture and cloud data
platform, here are the top 10 AWS announcements that excited me the most. Do
these resonate with you?
This is probably the most exciting announcement of the conference. AWS's
commitment to the Apache Iceberg open table format and the introduction of S3
Tables leveraging Parquet as the underlying storage will quickly enable
interoperability among AWS and popular cloud data platforms like Databricks and
Snowflake. However, be cautious as compute-specific optimizations may differ
among vendors.
S3 Metadata provides a quick way to establish data governance for S3 objects.
It's great that S3 makers themselves are providing this info, potentially as a
side effect of having S3 tables. Real-time business use cases leveraging
metadata could be interesting.
AWS is creating its own Large Language Model (LLM) Foundation Models, promising
leading performance and low cost. While I'm optimistic about the possibilities,
it's too early to be excited about another LLM—testing is necessary.
Amazon Aurora is going global with an active-active configuration for
distributed multi-region SQL databases. This reminded me of GCP Cloud Spanner,
which is not new but very costly. Hopefully, AWS will offer a more
cost-effective alternative.
Modernizing .NET or mainframe legacy applications is notoriously challenging.
Let's see if Amazon Q Developer, with the latest Generative AI technology, can
simplify the monolith effort—even a 50% success rate would be huge.
Generative AI is enabling intelligent self-discovery and understanding of
existing data. It's good to see Amazon Q enabling QuickSight similarly. This is
a significant step forward in enabling self-serve reporting for business users
with limited SQL or programming skills.
4
During my time in the retail industry, there were always requirements to ensure
direct paths between cloud-to-cloud or on-prem-to-cloud services, particularly
for compliance. This could be a good alternative to using zone-segregated load
balancers.
With the new S3 Tables in Iceberg, change data capture from Aurora
Postgres/MySQL database transactions can now be exported for analytical purposes
in near real-time.
Multimodal data support is expected behavior in publicly available LLMs. It's
great to see this enhancement in Bedrock for AWS users.
Business workflows are complex, often involving multiple divisions. Enabling
multi-agent support distributes functions and allows specialization of agents.
We love sharing what we've learned with you! Let's chat over a virtual coffee
before the holidays. Wishing you a Happy and Exciting 2025!!! 🥳🎉🎊
https://rearc.io/contact