Reconfigurable Computing Breathes New Life into NFV

Though Software-Defined Networking (SDN) and Network Functions Virtualization (NFV) have been around for a while now, many organizations are still weighing the benefits and drawbacks of adopting these technologies. SDN and NFV hold great promise because they separate software from hardware, which eliminates standard proprietary bundling and its hefty price tag. However, they also have challenges that organizations need to overcome before enjoying the full value that is possible.

In the early days of network infrastructure, the standard practice was to buy customized hardware and software. Example applications include network gateways, switches, routers, network load balancers, varied mobile applications in the mobile core; radio access network such as vEPC (virtual evolved packet core), vCPE (virtual customer premise equipment) and vRAN (virtual Radio Access Network); and security applications like firewalls, NGFW, IDS/IPS, SSL/IPsec offload appliances, DLP and antivirus applications. 

This requires buying proprietary appliances to run each networking application. Operators would rather support these functions as software applications (virtualized network functions, or VNFs), running on virtual machines or in containers on standard servers.

That’s the idea behind NFV. Moving away from discrete, cus­tomized architectures to a more consolidated “x86-only architecture” promises to reduce costs, simplify deployment and management of net­working infrastructure, widen supplier choice and, ultimately, enable horizontal scale-out in the networking and security market.

Because the throughput and latency demands of today’s applications are so high, there’s no guarantee that applications in software on standard platforms will be able to meet those demands without allotting significant CPU resources to address the issue. Operators are realizing that the cost savings that NFV promises are offset by the need to deploy entire racks of compute resources at a problem that a single appliance could previously support. The CPU and server costs, rack space, and power required to meet the same performance footprint of a dedicated solution end up being as expensive as or more than custom-designed alternatives. The vision of operational simplicity and dramatically lower total cost of ownership are still a dream on the horizon.

How 5G Complicates Things

Operators are already facing performance and scaling problems with generic NFV infrastructure (NFVi), and as 5G networks become a working reality, their presence will only make the situation worse. The move to 5G brings new requirements to mobile networks, creating its own version of hyperscale networking that is needed to meet the performance goals for the technology but at the right economy of scale. Numerous factors are fundamentally unique to 5G networks when compared to previous 3G/4G instantiations of mobile protocols. The shorter the distance, the higher the frequency – thus, the more bandwidth that can be driven over the wireless network.

On top of this, 5G will also mean a huge increase in the number of users/devices (both human and IoT), which fundamentally affects the number of unique flows in the network and necessitates very low latency requirements. 5G also promises lower energy and cost than previous mobile technologies. These 5G goals, when realized, will drive the application of wireless communications to completely new areas never seen before.

Acceleration Needed

Operators see now that in order to scale virtualized networking functions (VNFs) to meet performance goals, they will need data plane acceleration based on FPGA-based SmartNICs. This technique offloads the x86 processors that are hosting the varied VNFs to support the breadth of services promised.

It turns out that the highest-performing and most secure method of deploying VNFs involves virtual switching supported by SmartNIC acceleration. Virtual machines (VMs) can use accelerated packet I/O and guaranteed traffic isolation via hardware while maintaining vSwitch functionality. FPGA-based SmartNICs specialize in the match/action processing required for vSwitches and can offload critical security processing, freeing up CPU resources for VNF applications. Functions like virtual switching, flow classification, filtering, intelligent load balancing, and encryption/decryption can all be performed in the SmartNIC and offloaded from the x86 processor housing the VNFs while, through technologies like VirtIO, be transparent to the VNF, providing a common management and orchestration layer to the network fabric.

In addition, custom offloads beyond the above examples that are specific to the VNF application in question and that require acceleration can be implemented in the FPGA SmartNIC via standard APIs. This level of complete flexibility provides a workload-specific processing architecture where specific tasks are split between the host x86 processor and the FPGA.

A New Configuration

These changes are such that costly, hardened networking and security solutions simply will not suffice. The technique to overcome the challenges that are facing NFV deployments requires reconfigurable computing platforms based on standard servers capable of offloading and accelerating compute-intensive workloads, either in an inline or look-aside model to appropriately distribute workloads between x86 general-purpose processors and software-reconfigurable, FPGA-based SmartNICs optimized for virtualized environments.

By combining commodity server platforms and FPGA-based SmartNICs, the stage is set for an environment in which network applications can operate at hundreds of gigabits of throughput with support for many millions of simultaneous flows. Organizations that have been hanging back to see if the promise of NFV would become a reality can begin to build this unique architecture for networking applications.