AWS re:Invent re:Cap Pt. 2

Data. Where is my data? Where can my data be? How do I get my data there? How do I access my data? What kind of data can I use?

In recent years as Vendor 1 and Vendor 2 have started to pick up market share in the IaaS space, but the reason AWS will continue to be the best choice for enterprises is AWS’s features and functionality for Database PaaS. AWS is the only cloud provider that enables us with fully managed database platforms for Oracle DB (Vendor 4 if you are keeping score), SQL Server, MySQL, MariaDB, PostgreSQL. Amazon RDS also features Aurora which answers the need for highly transactional databases using Database engines MySQL or PostgreSQL.

At re:Invent 2017, AWS announced innovation in the database space that will keep AWS out front for years to come.

Amazon DynamoDB Global Tables

About 6 months ago I had a discussion with the Chief Architect at one of my favorite customers on the topic of the merits of Vendor 2 solution for NoSQL databases, Cloud DataStore. He finally said, “none of it matters because what I really need is to be able to read and write from anywhere and I can’t deal with the complexity of synchronizing between defined termination points. When are we going to get a global database as a service?” Well, we now have it.

DynamoDB Global Tables is fully managed and allows us to read and write to the endpoint that is geographically closest to our customer. Not only does this service give our customers greater performance through geo-dispersed access, it also gives them data resiliency (built in DR) all while we achieve five nines of availability.

Amazon Aurora Serverless (preview)

Even in modern architectures, when we have an RDBMS database involved, the database must always be available. The database has to be scaled to maximum demand. For this reason, in most architectures, the database is the most expensive infrastructure in the architecture. In addition to this cost, we often duplicate this cost in non-prod environments, at least doubling the price.

Aurora Serverless solves these problems with an on-demand database. Aurora Serverless will allow us to run databases that shut-down and cost nothing when our applications are not in use and automatically scale to meet demand as necessary. Database on-demand, pay only for what you use.

Amazon Aurora Multi-Master (preview)

Since the release of RDS (Relational Database Service), we’ve been able to make our data more readily available using two availability zones. We can read from multiple availability zones and we can fail-over to a secondary write node in the case of a failure. This has always been exciting because the functionality gave us similar cluster capabilities available from Vendor 1 and Vendor 4 all as a fully managed service. But what no one could do was active-active multi-master write to a scalable, performant, highly transactional database.

Aurora Multi-Master provides multi-master write across availability zones making our database architectures more resilient than ever. This is exciting because the combination of already existing functionality plus the availability of Multi-Master write capabilities takes the transformation of databases from Vendor 4 to Amazon Aurora beyond cost and performance into reliability, resiliency, and availability. It’s time to make the move. The product is simply superior.

Amazon Neptune (preview)

Neptune is a graph database available as a fully managed service. Neptune is fast, reliable, continuously backs itself up and is a high performant. Graph databases store relationships of structured and unstructured data. So, I can build a model that will allow me to query the relationships between the first call our sales team had with my customer (audio), the other people who participated in the process of sales and delivery, the technologies we delivered, the stats of the project and the picture of the look on my customer's face when they realized they saved 4 million dollars thanks to their move from Vendor 4 to Amazon Aurora.

PS. In Andy Jassy’s Keynote he avoided using the names of the competition. Instead, he just colored their Vendor number as I have here. For those who may not be able to discern who those Vendors are: Vendor 1 = Microsoft Azure, Vendor 2 = Google Cloud Platform, Vendor 4 = Oracle.