Accelerating the development of edge computing

5G is being rolled out faster than any previous generation of wireless technology. Research from Omdia shows that 5G subscriptions grew fourfold in 1Q20 to over 64 million and are on track to reach 10 billion by the end of 2024. In terms of subscribers, this is four times the adoption rate of 4G.

As 5G implementations accelerate, they bring billions of new connected devices, petabytes of additional network traffic, and millions of new 5G base stations, of which there are currently about 70 million base station locations worldwide.

Most said they expect to double or triple their number of antennas over the next five years. That would equate to nearly 150 million new base stations over the next five years.

A matter of scale

How do operators plan for the rapid expansion of base stations? How do they reduce costs (OPEX and CAPEX), minimize power consumption, deploy zero-touch solutions in light-free data centers, and reduce hardware SKUs and configurations? As a first step, operators are virtualizing large parts of their operations. They start with the 4G core and move to the radio access network - nearly 80% of which will be virtualized by the end of 2024, according to Omdia.

Source: Omdia

However, operators are still struggling with the question of how to deploy a highly distributed edge that could number hundreds of thousands, or even more. This distributed edge must be affordable and must maintain the flexibility and agility of virtualized, cloud-native implementations. Edge computing locations must also be able to support the applications that 5G is designed around: enhanced mobile broadband (eMBB), massive machine-to-machine communications (mMTC), and ultra-reliable low-latency communications (URLLC). While 5G implementations are initially focused on eMBB, network requirements for all three application types are growing rapidly.

Source: Heavy Reading

Expanding Solutions with FPGAs

The advantage of eMBB platforms and commercial applications is low cost. Therefore, operators are expanding their commercial off-the-shelf, software-only solutions with field programmable gate arrays (FPGAs) optimized for networking and security functions. These customized platforms leverage (FPGA) technology to achieve greater flexibility, agility, and scalability. They complement software solutions with higher performance and lower latency while freeing up limited and expensive CPU cores.

To meet performance targets for virtual network functions (VNFs), operators have realized that data plane acceleration (most likely through FPGA-based SmartNICs) is required. FPGA-based acceleration solutions enable operators to achieve higher speeds and lower latency while reducing power consumption and server-based capital expenditures by reducing the number of CPU cores required.

Some networks are well suited to running on x86-based CPUs. Others test monopolize general-purpose processors, wasting capital expenditures, floor space, and power in the process. Workloads that benefit from acceleration and offloading, especially at the network edge, are network- and security-related tasks such as switching, routing, operations processing, flow management, load balancing, and encryption. Example workloads include Open vSwitch (OVS), 5G/User Plane Function (UPF), Segment Routing Header (SRH), Key-Value Store (KVS), and overlay network tunneling protocols such as VXLAN and Network Virtualization using Generic Routing Encapsulation (NVGRE).

The intersection of performance, cost, and standardization

The use of specialized network hardware devices in carrier networks is finite. However, plain old COTS servers, such as plain old telephone service (POTS), are also at risk. Leveraging COTS servers and FPGAs gives carriers the industry-standard platform they want and the high performance they must have.