Google is making its in-house software-defined networking technology available to organizations using its Compute Engine cloud computing platform.
Google has been at the forefront of software-defined networking (SDN) for the past couple of years, developing an OpenFlow-based system dubbed Andromeda for internal use. Now the Web giant is bringing that network virtualization technology to Compute Engine, a move that will enable customers to “see major performance gains in throughput over our already fast network connections,” Amin Vahdat, a distinguished engineer at Google, wrote in an April 2 post on the company’s blog.
Andromeda already is available in two Compute Engine zones—us-central1-b and Europe-west1-a. Over the next few months, Google will migrate all of its zones to the SDN platform, wrote Vahdat, who gave a presentation about Andromeda in March at the Open Networking Summit.
SDN and network-functions virtualization (NFV) have become the hot topics in a networking market that is seeing growing demand for infrastructures that can scale and are more automated and programmable than the traditional hardware-based environments. SDN essentially removes the network intelligence from the underlying physical infrastructure—such as complex and expensive switches and routers—and houses it in software-based controllers. Virtualized networking functions—from load balancing to firewalls—are available as software applications.
For big Web-based companies like Google, Facebook, Microsoft and Amazon, being able to scale rapidly and adapt to changing business demands is important. In addition, in a business world that is changing due to such trends as cloud computing, big data, virtualization, greater IT mobility and bring-your-own-device (BYOD), organizations are seeing a growing need for networks that are more automated, easier to program and less costly.
Essentially all major networking vendors are building out their SDN and NFV portfolios and defining what they say are the best ways to implement SDN. Cisco Systems is taking a different tack with its Application Centric Infrastructure (ACI), which offers more of a software-and-hardware approach to achieve similar goals.
Google has been aggressive in building much of its own data center resources, including its networking. That gives the company an edge in orchestrating everything from virtual machines and hypervisors to operating systems, networking gear and network interface cards, Vadhat wrote.
“We are uniquely positioned to leverage Google’s control and expertise over the entire hardware, software, LAN, and WAN to deliver a seamless experience for Cloud Platform customers,” he wrote. “At Google, we benefit from having programmable access to the entire network stack, from the lowest-level hardware to the highest-level software elements. Rather than being forced to create compromised solutions based on available insertion points, we can design end-to-end secure and performant solutions by coordinating across the stack.”
Andromeda helps ramp up the performance of the network while exposing NFV, including such functions as protection against distributed denial-of-service (DDoS) attacks, load balancing, access control lists and firewalls, he said.
“Hence, Andromeda itself is not a Cloud Platform networking product; rather, it is the basis for delivering Cloud Platform networking services with high performance, availability, isolation, and security,” he wrote.
In addition, Google engineers already are working on improvements that will include enabling high-speed access to low-latency, durable storage, APIs for NVF and virtual machine migration.
“Andromeda is a re-working of our underlying network virtualization architecture, and its SDN core enables us to rapidly iterate and deliver new functionality,” Vadhat wrote. “This ensures that Cloud Platform’s network will continue to be an agent of disruption to cloud computing moving forward.”