Amazon Web Services is offering up its latest general-purpose virtual servers that come with a powerful custom Intel processor and networking technology that can reduce latency between instances by as much as 50 percent or more.
AWS officials on June 11 announced the M4 instances for their Elastic Compute Cloud (EC2), which also include dedicated bandwidth to Amazon's Elastic Block Store (EBS) for storage. The world's top cloud provider is offering five versions of the latest instance, with virtual CPU count ranging from two to 40, memory from 8GB to 160GB, and network performance as high as 10G bps.
"The supersized core count on [the m4.10xlarge] instance [with 40 virtual CPUs] will be great for applications that use multiple processes to achieve a high degree of concurrency," Jeff Barr, chief evangelist for AWS, wrote in a post on the company blog.
AWS officials said the new instances—or virtual servers—can handle a wide range of applications, including relational and in-memory databases, gaming servers, batch processing and business applications like SAP and Microsoft's SharePoint.
Intel over the past couple of years has been building out its custom chip business to serve cloud and Web-scale operations. During the company's latest quarterly earnings call in April, CEO Brian Krzanich noted that the newest Xeon E5 "Grantley" chips and "our increased support of custom versions of the product helped us achieve record cloud revenue [in the first quarter]." Most recently, Intel last month announced a partnership with eASIC to bring application-specific ICs (ASICs) to custom Xeons that can be used in enterprise data centers and cloud environments. The move is part of a larger push by the giant chip maker to expand its custom chip efforts by leveraging accelerators like field-programmable gate arrays (FPGAs) and ASICs for particular workloads.
The chips used in the latest AWS instances are custom Xeon E5-2676 v3 "Haswell" chips optimized for EC2, according to Barr. The chips have a base speed of 2.4GHz, but they can crank as high as 3GHz via Intel's Turbo Boost technology.
He also pointed to the Enhanced Networking technology that comes with the M4 instances that provide up to four times the packet rate of those instances without the technology, noting that Enhanced Networking can help reduce latencies between instances by as much as 50 percent. It also helps reduce jitter, AWS officials said. The M4s are the first general-purpose instances to include the Enhanced Networking feature, according to Matt Garman, vice president of EC2.
"With these capabilities, M4 is one of our most powerful instance types and a terrific choice for workloads requiring a balance of compute, memory, and network resources," Garman said in a statement.
The M4 instances are available in various regions in the United States, Europe and Asia Pacific. They can be launched in On-Demand or Spot form. In addition, Reserved Instances can be bought, Barr said.
Along with the new M4 instances, AWS is reducing the On-Demand and Spot pricing for its M3 and C4 instances by 5 percent in U.S. East (Northern Virginia) and West (Oregon) regions, Europe (Ireland and Frankfurt, Germany), and the Asia-Pacific region (Tokyo and Sydney, Australia). The price reductions are available now.