Packet expands its global infrastructure with 11 new edge locations
Packet is quickly making a name for itself thanks to its unopinionated bare metal cloud. Today, the company is expanding its product portfolio with the launch of its new Edge Compute service in 15 data center locations around the world. Eleven of these locations are new and complement Packet’s existing offerings in New York, Sunnyvale, Amsterdam and Tokyo.
The idea behind these new edge locations, which are now online in Los Angeles, Seattle, Dallas, Chicago, Ashburn, Atlanta, Toronto, Frankfurt, Singapore, Hong Kong and Sydney (with Paris, London, Sao Paulo, and Mumbai coming later this year), is to give Packet’s users the ability to easily deploy a global low-latency infrastructure. Unlike its regular locations, these new edge locations will currently only feature a single, Intel SkyLake-based server configuration, though the plan is to bring most of Packet’s server options to these locations over time.
The company argues that while content delivery networks allow you to keep your content close to the edge of the network, a lot of businesses now also want to do computation close to where their users are. To do that, a company like Packet needs to be able to serve its user base where it is.
As it adds more server configurations and expands these edge locations, one could argue that what is an edge location today simply becomes another data center for Packet in the long run. Packet co-founder and CEO Zachary Smith, however, argues that this isn’t the plan. “Although some sites may evolve to offer our complete computing portfolio, our core goal will be to reach even more locations and bring fundamental infrastructure even closer to latency-sensitive applications and users,” he told me. “These edge sites will serve as a unique test bed for developers to hone the pricing, contracting, automation and capacity management techniques they will need as infrastructure becomes even more distributed.”
Smith believes that Packet’s advantage over other platforms is that it gives developers a completely unopinionated platform that gives them full control over the infrastructure. He also noted that as developers look to building IoT solutions, autonomous cars, and similar data-heavy products, they will need a global network of data centers that are close to their users. “In the future we expect the number of locations to grow and the diversity of hardware to evolve,” he says in today’s announcement. “But you won’t see us opening up 1,000 football sized compute farms – those thousand locations will be look a lot more like the hipster specialty food market on the corner of your block, instead of the Sam’s Club 30 minutes away near the freeway.”
Looking ahead, Packet also believes that this kind of next-generation cloud computing will need specialized hardware. Google, Microsoft and others are already deploying custom chips in their data centers and the Packet team believes that there is room for a service that gives developers access to a similarly diverse set of machines.
Featured Image: Getty Images