‘Speed is just a question of money: how fast do you want to go?’ is an old hot-rodder’s saying. And it also applies to telcos and MNOs. They have invested heavily in their 4G and LTE networks, embracing virtualization to enable faster, more agile and scalable deployments that can keep pace with the explosion in subscriber data traffic and consumption, while keeping control of their overall CAPEX and OPEX costs. Even though these investments have already given us mobile connectivity of unprecedented speed and pervasiveness, they have only laid the foundation of what’s to come, as 5G and smart connected devices start to roll out.
A recent study by a division of Nokia Bell Labs* gave a glimpse of what’s coming up: it found that the number of Internet of Things (IoT) connected devices is expected to expand from 1.6 billion in 2014 to well over 20 billion by 2020. Also, by 2020 global consumption demand for digital content and services on portable devices will see an average increase of 30 to 45 times the levels seen in 2014. So mobile operators will need to further accelerate their technology investments to meet ever-increasing consumer and business connectivity demands.
The IoT explosion
Although mobile operators in the US are starting to embrace ‘pre-standard’ 5G, vying for the first mover’s advantage, the standards are not actually expected until 2018. This has echoes of the early days of 3G in the late nineties and early 2000s, when proprietary implementations were later standardized. Unsurprising, then, that some operators in the US have already announced their plans to deploy 5G later this year, and some of the early adopters in Korea, Japan and China are ready to roll out. As such, many 5G concepts are being solidified fast.
5G is sometimes narrowly classified as a higher bandwidth radio access technology. But it is much more than that. It will also be the network for low-power devices and sensors that are classified as IoT devices, as well as low latency applications. One example is in regards to the low latency for some mission-critical devices such as connected, driverless cars. This will dictate that key LTE base station functions are distributed, with some moving closer to the edge and the others being ‘pooled’ in the cloud.
Another example is that vertical sectors will require different types of services from the 5G network. Some will need high bandwidth, and others low power. Some will need very low latency, others very high availability. The vast volume of IoT devices will range from those that send multiple Gbps to those that will send a few bits every month. This flexibility and elasticity can only be supported by advanced network virtualization.
Focus on the service
In order to support such varying 5G use cases across multiple verticals, operators need to shift from being network centric to being more service oriented.
To understand this shift, let’s examine the concept of ‘network slicing’. Consider cases such as mobile broadband with higher bandwidth video requiring high availability everywhere, or low power sensors that can operate on a pair of AA batteries for 10 years, or connected driverless cars that can zip through crowded streets in Los Angeles or Seoul but brake within microseconds when they sense potential obstructions in their paths. Each of these requires a different configuration of the requirements and parameters in the network. In essence, each needs its own network slice.
As such, networks need to be built such that speed, availability, capacity and coverage can all be allocated in logical slices to meet specific demands of these use cases.
The best way to implement these network slices will be via virtualization – to use service provider SDN, NFV and network orchestration. In each of these cases, the SDN controller will configure and build network slices that include service chains such as DPI (Deep Packet Inspection), emails, and security scans per user, or per service. The network services themselves are virtualized as Virtual Network Functions (VNFs) that allow operators to set up services rapidly, and scale them in response to network and service demands.
Managing the migration
We can see, then, that virtualization will be a vital piece of the puzzle as 5G is rolled out and the IoT grows. But the speed and agility that virtualization enables must be balanced against the need for network visibility, resilience, manageability and security, throughout the development, roll-out and mass usage of each service.
MNOs and service providers must weigh the trade-offs between quality and cost; flexibility and control, between moving too quickly, and not moving fast enough. As a result, the demand for simple, end-to-end solutions that can efficiently test and validate the effectiveness and benefits of virtualization at every stage is only going to increase. Headlines about the Internet of things are still dominated by the exciting consumer stories – intelligent fridges and cars that can drive themselves. But network virtualization will be a key driver of the IoT ecosystem as it develops – and full-lifecycle reliability and performance testing will be a key element of this shift.