Share, , Google Plus, Pinterest,

Print

Posted in:

Mobile Edge Computing (MEC) – A Step Towards 5G

Where is the edge of a mobile network? This obvious question comes to mind when we first think of the topic “Mobile Edge Computing” (MEC). If we look at the mobile network architecture, the radio and core network elements in a mobile network are the places where a lot of sophisticated signal processing happens and any application server would be sitting near the core network. So, the edge can be assumed to be somewhere between the IP network and the beginning of the mobile network i.e. the core network. MEC is changing this by bringing traditional IT infrastructure and applications deep into the mobile network, all the way to the radio network elements. MEC is a new standardization initiative supported by market leaders like Nokia, Huawei, NTT Docomo and Vodafone to name a few. Let’s try to look at why edge computing is needed and how it adds value to different stakeholders in the industry.

Over the past many years, Internet Protocol (IP) has spread from the internet to the enterprise servers and thereafter to personal computers. With widespread adoption of mobile technologies like 3G and LTE, the IP has now moved all the way through to the end mobile devices themselves. People now increasingly chat with friends and family, always stay active on social networking sites, watch movies and videos on the move, listen to streamed music and make video calls through their mobile phones. Mobile data traffic continues to grow at a fast pace due to the emergence of new applications. Gartner research predicts that mobile data traffic will grow at a rate of more than 50% in 2015-16 with a key driver for mobile data growth on a global scale being mobile apps, particularly mobile video apps.

Although the mobile data traffic is growing, the eroding ARPU (average revenue per user) is putting pressure on the network operators to optimize operations and resource utilization, come up with innovative revenue generating services and doing all these while maintaining the quality of service. Today’s businesses and the end users are demanding more from the mobile telecommunication industry. The end users want personalized services, better performance and better user experience. Imagine your shopping experience when an augmented reality solution would guide you through each floor in a mall providing personalized information on products and offers at different stores of your interest; or watching live content of a game on your mobile from different camera angles while you are in the stadium; or think of an application that provides you instant information on the different things on display in a museum just by placing the camera of your phone in front of the object. These are the kind of services that will become new revenue streams for an operator in the future. With rich quality of experience and greater customer satisfaction, they can help an operator retain existing customers and also attract new customers.

With the growing popularity of cloud computing, most of the smart phone applications and content are now gradually moving to the cloud because of the large pool of resources and services that the cloud facilitates. However, a centralized cloud data centre or a cloud platform brings along with it issues like unreliable latency, lack of location awareness etc. The solution examples mentioned above need network characteristics of very low latency, location awareness, context awareness, local or geo-distributed processing while still using the benefits of virtualization and cloud. In fact, end-to-end network latency of 1ms to 5ms is one of the key requirements for 5G aimed at a class of applications called the “Tactile Internet”, which is simply not feasible with traditional clouds. The billions of devices that will get connected with Internet of Things (IoT) will further congest the network and the network operators need to offload some amount of processing from the centralized data centres by doing local processing and data analysis wherever possible.

Mobile Edge Computing (MEC) enables cloud servers to run very close to the mobile devices (e.g. at a base station or an eNodeB), thus shifting a lot of computational effort from the centralized cloud to the mobile network’s edge and hence providing reduced latency and local processing of data, that will make real time applications like cloud gaming, augmented reality and real time video analytics possible. Edge clouds are not going to replace traditional data centres, as the amount of processing power and storage that they have is in orders of magnitude far below the traditional clouds. Instead, the goal is to complement the traditional cloud data centres by running some components at the edge of the network for delay sensitive applications.

Some of the current market ready solutions include Nokia Networks’ MEC platform “Radio Application Cloud Servers” that can be co-hosted with a base station which will allow the integration of lightweight third-party software components, extending the base stations’ core functionalities with application-specific features. Or Saguna’s “Open-RAN” MEC platform that operates inside the RAN and with its fully virtualized software architecture. A network operator can implement differentiated services at the edge. ADLINK’s new “Extreme Outdoor Server” is also a high-performance MEC platform specifically designed for extreme environments and outdoor telecom/networking applications.

With its potential to offer ultra-low latency, high bandwidth, traffic offloading capabilities, data analysis, aggregation and augmentation at the edge, MEC will open new frontiers for network operators, application service providers and content providers by enabling them to introduce innovative services and offer an ultimate user experience. MEC can also enable disruptive services in the area of IoT/M2M and Big Data analytics with its ability to connect, collect and analyse data at the radio network edge. However, the mobile edge network elements are typically far away from IT experts and may even be installed at remote locations. This means the edge computing platforms must be easy to install and configure, easy to upgrade, easy to use and manage, easy to monitor and most importantly they must be highly resilient and fault tolerant.