Amazon has announced a new addition to its EC2 lineup. The X2gd instances offer twice as much memory per virtual processor as their predecessors.
X2gd is intended to be the successor to the existing R6g instances. These EC2 instances are based on Amazon’s own Arm-based Graviton processors. Amazon has a wide range of these, optimised for different tasks, but R6g was intended for applications that required a lot of memory. Examples of such applications are Redis, open source relational databases, real-time analytics, caching services and containers.
Higher performance and lower costs
Compared to R6g, the new X2gd instances offer twice as much memory per virtual processor. In its press release, Amazon also likes to compare X2dg with its x86 counterpart, called X1. X2gd could, in fact, offer the same amount of memory per virtual processor as X1 but have a 55 percent better price/performance ratio.
“Graviton2 processors have proven incredibly popular with customers because they deliver a major improvement in performance while also significantly lowering costs for a wide variety of workloads,” said David Brown, Vice President, Amazon EC2, at AWS. “With up to 55% better price/performance than current generation X1 and double the memory per vCPU compared to the latest Graviton2 R6g instances, new X2gd instances provide the lowest cost per gigabyte of memory we have ever offered in EC2. X2gd instances give customers the ability to consolidate their memory-intensive workloads on instances with fewer vCPUs and realize game-changing performance benefits and cost savings.”
AWS is betting big on ARM. The company introduced its first generation of Graviton processors in 2018, only to follow it up a year later with the much more capable Graviton2 processors. AWS is now using these extensively for its EC2 instances. In May last year, the company announced the M6g, C6g and R6g instances. According to AWS, the instances are also eagerly purchased by parties such as Domo, Formula One, Honeycombi.io, Twitter and many other companies. Amazon itself also uses the instances for many of its cloud services.