S3? That’s so 2006. AWS deepens data infrastructure play

S3? That’s so 2006. AWS deepens data infrastructure play

AWS has come a long way from the days of its simple S3 object store, with increasingly sophisticated services to make it an indispensable piece of your data infrastructure.

At re:Invent this week, AWS went further – on a number of fronts.

We had two blockchain announcements. The Quantum Ledger Database lets you track and verify the complete history of changes to application data using an immutable journal that stores changes sequentially and uses cryptography to seal them.

QLDB ties in with the AWS Managed Blockchain service Amazon also launched and lets you deploy your own blockchains. That service will write directly to the Quantum Ledger, which runs up to three times faster than other blockchain frameworks, according to Amazon.

This development consolidates the AWS Blockchain Templates the company made in April. These let you deploy Ethereum and Hyperledger frameworks via AWS CloudFormation templates. Now, Managed Blockchain will support the same frameworks but will manage certificates and operational metrics automatically.

Another industry meta trend is IoT, so – naturally – AWS had something there at re:Invent, with Timestream. This is a managed, time-series database that targets those building IoT and industrial telemetry but that can also be deployed for application monitoring.

Timestream is a streaming data processing service, that takes trillions of events per day in time order and compresses them so users can later interrogate them via the product’s query engine. According to AWS, Timestream lets you store and analyse log data for DevOps along with sensor and telemetry data from IoT and industrial equipment. It will understand locations and format, automate rollouts, retention, tiering and data compression – all in the name of simplified management.

This service is, of course, serverless – demonstrating, once again, how AWS is moving increasingly away from services that expose server management. This lets you concentrate on just using the software but – of course – ties you ever more closely to Amazon’s cloud.

Cloudy AWS raised eyebrows when it announced a partnership with on-prem virty king VMware in 2016. This summer they announced Relational Database Service (RDS) on VMware but the service went live this week, at re:Invent. RDS on VMware takes AWS’s cloud-based RDS offering and lets you implement it on your servers, enabling the same management tools for on-premises systems as you’d use in the cloud. It also lets you build data recovery systems that interact with cloud-based RDS databases.

Through one management interface, you should be able to take care of automatic database provisioning, operating system and database patching, backup, point-in-time restore, storage and compute scaling, instance health monitoring and failover. Amazon RDS works with Amazon Aurora, PostgreSQL, MySQL, MariaDB, Oracle and SQL Server.

It was 2012 when AWS signalled its intention to move into the corporate data fabric with Glacier, a data back-up and archival service designed to kill off tape. This week, AWS unveiled Glacier Deep Archive – a new S3 storage class for archiving long-term data at low costs.

Amazon is still gunning for the archival tape market, offering $0.00099 per GB a month it says are comparable to tape archival services. Data can be retrieved in 12 hours or less, while a bulk retrieval option will let you pull back petabytes of data within 48 hours.

It’s another mechanism that Amazon can use to slurp large amounts of enterprise data into its cloud and drip feed profits.

Finally, the company announced a preview of AWS Lake Formation for big data. The service builds data lakes by crawling a range of data sources, using Machine Learning to de-duplicate the data, and then pulling it into an S3 bucket. It then optimises the data for analytical processing, and lets customers define security policies across the whole data lake.