Amazon Web Services (AWS) made it abundantly clear this week that the cloud service provider intends to aggressively compete as a provider of managed networking services. This follows the launch of AWS Global Accelerator, a managed service based on the global network the cloud service provider built to connect 19 availability zones that collectively house well over a hundred massive data centers.
The service was announced at the AWS re:Invent 2018 conference, and runs across a global network spanning over 150 points of presence (PoPs), in addition to 89 AWS Direct Connect locations through which IT organizations can physically connect to the AWS network. According to Peter Desantis, vice president of global infrastructure and customer support for AWS, who spoke during an opening keynote yesterday, the network is comprised of both the most dense instance of fiber optic cable and undersea water cables available anywhere.
“This is a purpose-built cloud network,” said Desantis.
As part of the strategy, AWS also announced an AWS Transit Gateway through which organizations can set up a traditional hub and spoke network topology using AWS networks and an Elastic Fabric Adapter. This will allow it to become what Desantis claims is the first cloud service provider to support 100G Ethernet. There’s also a new C5n offering based on switches and network virtualization software that AWS developed to make additional network bandwidth available for latency sensitive applications, a lightweight virtual machine — dubbed Firecracker — that is optimized for running containers, and several extensions to AWS Internet of Things (IoT) services.
Finally, AWS also announced it has added A1 instances based on AWS Graviton processors developed using the Arm architecture, along with a P3dn GPU instance optimized for distributed machine learning applications. Desantis noted that AWS can add all these instances of compute, networking, and storage services because it developed a Nitro system that offloads the management of those services from the core processors made available on the AWS cloud. Desantis explains that this approach not only makes it simpler to add services, it frees up 100 percent of the capacity of those processors for applications.
The challenge and opportunity for MSPs
The rise of AWS as a provider of managed networking services cuts both ways for MSPs. Those that choose to partner with AWS now have access to a global network that is managed by a single entity around the world. AWS partners don’t have to invest in physical switches either. Those MSPs that don’t partner with AWS will continue to rely on network connections spanning multiple carriers that don’t always provide each other with much visibility into their respective networks. Desantis made it clear AWS is betting most enterprise IT organizations are going to prefer a global network that is managed by a single provider.
AWS is betting most enterprise IT organizations are going to prefer a global network that is managed by a single provider.
Of course, not every application an MSP is asked to manage is going to reside inside an AWS availability zone. But there is now enough critical mass of applications residing in the AWS that it has become hard to ignore. In fact, AWS networking services give MSPs some leverage against telecommunications carriers that are not especially anxious to see more networking traffic shift over to a network they don’t own or manage. The challenge and opportunity facing MSPs now is figuring out how best to leverage a wide range of expanding AWS networking services to their best advantage.
Photo: Sundry Photography/Shutterstock