Mobile Edge Computing (MEC): Base Station vs. Aggregation Point


When it comes to Mobile Edge Computing the first question that comes to mind is: Where is the edge?
For a long time, the edge in mobile networks defined the boundary between the IP network and the beginning of the mobile network; the core network. Mobile edge computing (MEC) is changing this, bringing traditional IT infrastructure deep into the mobile network, all the way to the radio access network (RAN).

Mobile Edge Computing

Mobile Edge Computing is a new ETSI standardization initiative, supported by market leaders. It is an evolutionary step of NFV, which discusses the separation of functionality from the underlying hardware infrastructure to increase mobile network flexibility, economy and scale. Mobile Edge Computing (MEC) takes this concept one step further. By bringing computational power into the mobile RAN, MEC promotes virtualized software based ecosystems at the radio edge. The MEC platforms enable virtualized applications to run much closer to mobile users to boost the user experience.

The MEC approach enables new players in the market to enter the Radio Access Network, which has traditionally been monopolized by a select few giants in the mobile networking market space. It also opens a new mobile arena for software players and web services and content providers seeking to “get closer” to mobile users.

Base Station vs. Aggregation Point

Yet, the question we opened this post with, remains open: inside the mobile Radio Access Network, which is the preferred edge? Is it inside the base station or at an aggregation point? In the ETSI specification group, both alternatives are being discussed.

The ETSI MEC group has stated its goal “to flexibly and rapidly deploy innovative applications and services towards mobile subscribers, enterprises and vertical segments.” To achieve this and open the door to the widest possible number of players from network, IT and content providers, the MEC standard would need to support both Base Station platforms and aggregation based (Aggregation Points) platforms.

Let’s consider the two options.
The incentive for integrating the MEC platform inside the base station is obvious. This is the closest point in the mobile network to the mobile user and, therefore, the one with minimal latency. This use case was demonstrated Saguna and Akamai in their showcase of the first Content Delivery Network operating inside the base station in CTIA 2014. In the demonstration, Akamai’s Content Delivery Network (CDN) was integrated as a server-side application running on the Saguna’s platform, which operates inside the mobile base station.

On the other hand, when multiple base stations are located close together, it may be useful to consider locating the MEC platform at an aggregation point. Serving several base stations with a single MEC platform located at a central aggregation point within the RAN centralizes the resources and reduces the CAPEX and OPEX. This approach can be more cost effective without adding a significant amount of latency.

The challenge of locating the aggregation-point MEC platform is the MEC standard requirement to provide radio-network contextual information. Note that the MEC standard discussion calls for Radio Network Information Services (RNIS) to be provided as a service of the MEC server service. The radio network information is provided to MEC applications and web services. This information is readily available to the MEC platform residing inside a base station. However an aggregation-point MEC platform may be serving one or more base station that does not provide this information since they do not support the MEC standard. To comply with the proposed MEC standard, in such a case, the Aggregation-point MEC will need a new set of tools to calculate the congestion level and network status based on traffic and loads.