My New Car, the Rise of Multi-Access Edge Computing, and the BIG-IP i15000 Series

Glenn Graham

Published February 21, 2018

My new car. Lots of data.

I recently bought my first new car in 10 years. Along with the fact that it runs well, I love the technology and numbers that are accessible to me now. I’m a spreadsheet sort of person, so having these numbers is fun. I can see miles per gallon, miles left in the tank, percentage of electric vehicle driving ratio, etc. And soon, there will be even more data available to me, the driver. Connected cars are a high-profile new network application that 5G will enhance. Intelligent driving, maps with real time data, and driving assistance based on cloud computing are areas that require cars to be connected to the cloud and the network, according to the Automotive Edge Computing Consortium (AECC)1.

A need for a new architecture.

I particularly like the idea of an indicator if traffic ahead has stopped (which would have negated my need to buy a new car) and a warning when cars are coming into your lane, or the lane you’re moving into when overtaking—both are new applications the AECC suggests. These require that real time data be transmitted from both your vehicle and other drivers’ vehicles. The AECC forecast 100 million connected cars in 2025, with a total data volume of 100 petabytes per month, based on an assumption that each car generates 1 gigabit per month. Note: that figure is based on current applications, not future ones, so should be considered very conservative. Globally, M2M connections of the Internet of Things, will grow 2.4-fold, from 5.8 billion in 2016 to 13.7 billion by 2021. Connected cars will be one of the fastest growing segments at 29 percent CAGR, equal to connected cities applications and second only to connected healthcare at 30 percent2.

Enter multi-access edge computing.

This growth will result in a significant impact on the network and the AECC plans to investigate new technologies that fulfill the system requirements they’ve specified. One goal of the new network design is multi-access edge computing (MEC), which moves compute and storage functions away from the core of the network and closer to the end user, better accommodating traffic to and from a massive number of smart devices. The new network design must reduce both application latency and the amount of data traversing the transport network between the core and the server location.

However, with the increasing implementation of IoT applications, the need for MEC is already here. While it will be a cornerstone for 5G networks, it is just as important for evolving 4G technologies. What are the MEC drivers and dependencies today? We’ve mentioned reduced latency and data traffic, which will also reduce the cost of backhaul. But networks also need more security at the edge to localize vulnerability. Each of the millions of IoT decides on the network is an opportunity for a malicious attack. Networks will leverage virtualization and network slicing, in which the operator can ‘slice’ the network into different segments from the radio network to the core and allocate different levels of service to each slice.

The evolution of 4G networks.

The reality is, many of these things can already be done in 4G networks. F5 has a long history of application delivery expertise in both software and hardware, and is well placed to support the necessary evolution. DNS caching and IP/TCP optimization (both strengths of F5) will play a critical role. Network slicing is available through F5’s GTP (GPRS Tunneling Protocol) session director. F5 also offers a wide variety of virtualized network functions and has the broadest range of performance options—through to 40GB—and virtualization platforms in the industry.

And F5 has just announced the BIG-IP i15000 Series, its highest performing services delivery appliance. In a 2-rack unit form factor, it provides high-performance and scalability to support the transition from 4G to 5G networks, connect to the multitude of IoT devices, and ensure that service provider networks, data, and subscribers remain secure against DDoS attacks. The i15000 Series enables the consolidation of multiple services onto a single device: services like load balancing, IP/TCP optimization, firewall, DDoS protection, CGNAT, traffic steering, and DNS caching and security. With 320 Gbps of L4 throughput, the i15000 Series supports demanding service applications with 300M concurrent L4 connections (e.g. HTTP load balancing requests) and 4.5M new connections per second, with high-density, high-performance ports (4 x 100Gbps and 8 x 40 Gbps).

A 5G world is coming soon—F5 solutions can help as your 4G networks evolve. And to take advantage of what’s on the horizon, I shouldn’t wait another 10 years before buying my next car.

References