AWS re: Invent. The main announcements of the first day (Part 2)

4 min


We continue to publish announcements of new AWS products from the main cloud event of the year – AWS re: Invent. In the first part, we talked about all the key events presented by the visionary of the company Andy Jassy. You can read it here. Now the main changes announced in the field of data storage.

New type of Amazon EBS volumes – gp3

Next-generation Amazon EBS General Purpose Block Storage SSDs that enable customers to vary performance regardless of storage capacity and offer up to 20% lower price per GB than existing gp2 drives.

The new gp3 volumes deliver a baseline performance of 3,000 IOPS (IOPS) and 125 MB / s (throughput) at any volume size. Customers can scale up to 16,000 IOPS and 1,000 MB / s for an additional cost if an application requires better performance than baseline. You can easily migrate data from gp2 to gp3 using the existing Elastic volumes functionality.

Gp3 volumes are recommended for applications that require high performance at a reasonable cost, such as MySQL, Cassandra, VDI, or Hadoop clusters.

Details here

Tiered pricing for Amazon EBS io2 based on IOPS performance

The first 32,000 IOPS are charged at the current base rate. Additional performance redundancy in the range of 32,001-64,000 IOPS is charged with a 30% discount. As a result, it now costs $ 3,552 to reserve 64,000 IOPS per month, saving customers 15% of their costs. When using the io2 Block Express volume, performance over 64,000 IOPS is billed with another additional 30% discount. Customers currently using Amazon EBS io2 volumes with over 32,000 IOPS in Reserved will automatically benefit from the new pricing starting December 1, 2020. Tiered pricing applies only to individual Amazon EBS volumes, not the total IOPS of all reserved volumes io2.

More details here

Support for multi-attach options for Amazon EBS io2 volumes

The multi-attach option allows the Amazon EBS volume to be used simultaneously for writing and reading from multiple (up to 16) servers. This functionality is now supported for both io1 and io2 volumes. Amazon EC2 VMs must be Nitro managed and located in the same Availability Zone. Currently only Linux operating system is supported. The functionality does not require additional payment.

io2 Block Express (SAN in the cloud)

Some clients require extremely high performance in the hundreds of thousands of IOPS. Already, as part of the Preview, you can test the io2 disks that run on the new generation of the EBS Block Express block storage architecture. Io2 Block Express drives deliver peak performance of 256,000 IOPS and 4,000 MB / s with average latencies of less than a millisecond. The functionality of data encryption @ Rest and creation of snapshots is supported. Io2 Block Express volumes are suitable for Oracle, SAP HANA, Microsoft SQL Server, and SAS Analytics startup tasks.

Details here

Amazon EC2 R5b instances for optimal performance with io2 Block Express

Amazon EBS access performance on new Amazon EC2 R5b instances is 3x faster than previous generation R5. Instances are powered by next generation AWS Nitro cards. This allows, depending on the size of the R5b virtual machine, to achieve performance in the range of 10-60 Gbps and 43-260 IOPS. Amazon EC2 R5b is suitable for high-load relational databases and ERP system applications.

Details here

Amazon EBS Volume Cost Optimization Tips

Compute Optimizer analyzes io2 / io2 and gp2 / gp3 volumes settings and provides recommendations for optimizing their performance. The functionality does not require additional payment.

Optimizing @ Rest encryption for Amazon S3

The new ability to manage SSE-KMS encryption keys at the S3 bucket level significantly improves the internal performance of @Rest encryption. By eliminating the need for ongoing calls to AWS KMS Key Management, the cost of encryption is 99% lower.

Replicating Amazon S3 data to multiple buckets

The new functionality enables you to simultaneously replicate Amazon S3 data to multiple buckets located in one or more AWS Regions. This makes it easier to build geographically dispersed solutions. The main application is to reduce latency when accessing Amazon S3 resources from different geographic locations, as well as building distributed disaster recovery solutions.

Strong consistency of Amazon S3 resources

Until now, Amazon S3 has provided read-after-write consistency when creating new objects. Eventual consistency was guaranteed when objects were modified or deleted. The new functionality ensures strong consistency of data in Amazon S3 buckets for all operations. This greatly simplifies the work with the data lake and analytical applications.

Russian-language Twitch session

And we continue to recommend the AWS Russian-language twitch stream, which will take place on the key days of AWS re: Invent. Streams are prepared and conducted by leading AWS solution architects who have chosen all the most interesting and useful from the news and announcements of the multi-hour conference. Register, connect to streams and discuss live.

For those who are late today or want to review twitch, where they discussed storage in detail, we post recording


0 Comments

Leave a Reply