Mobile communication providers are facing increasing competition from OTT players while simultaneously battling diminishing ARPU (average revenue per user). In addition, there is the added pressure of having to provide high bandwidth services and moving to 5G networks. The recent McKinsey report[i] 鈥� 鈥淭he road to 5G: The inevitable growth of infrastructure cost鈥�, elaborates on the dilemma faced by mobile operators. On the one hand, there is a need to launch 5G networks to address growing traffic and service quality requirements demanded by new applications such as AR/VR and gaming. On the other hand, 5G transmission uses high frequency bands with decreased propagation distances and requires heavy traffic processing, thereby calling for a substantial investment in radios, base stations and associated infrastructures to cover a sizeable geographic region.

The level of required investments is quite high, which makes things difficult for an individual operator. The high costs along with the perceived importance of 5G has some governments pondering over the idea of a government-created 5G network that could be shared amongst operators[ii]. While this idea seems far-fetched, operators have already slowly started sharing non-differentiating parts of the network, such as towers, antennas, base station, fiber, and electricity distribution poles. This blog proposes sharing the Multi-Access Edge Compute (MEC), an integral part of a 5G network infrastructure, which would take this sharing a step forward.

MEC is a new element defined in 5G architecture to meet promised operational and performance requirements. MEC integrates NFV-based cloud computing environments with radio access networks to not only provide the benefits of aggregation and statistical gains of shared resources, but also allow better response time and throughput. Many 5G services desire to run closer to the user due to one or more of the following reasons:

1. Performance: Propagation latency could cause problems
2. Business: Data transmission is too expensive and should be processed locally
3. Legal and security: Data should not leave the premises

Due to its importance, ETSI (The European Telecommunications Standards Institute) and other standard organizations have defined reference architecture for the MEC[iii]. In their framework, MEC consists of a mobile host, a virtualization or containerization infrastructure, and supporting services such as DNS and traffic rule policies, all working under the control of a platform manager. The platform manager allows running multiple mobile edge applications under the guidance and support of a centralized or distributed operations support system and orchestration. ETSI has also envisioned operating the MEC with various cloud services in order to seamlessly facilitate running applications in either the MEC or cloud. While this is a good step toward running multiple applications on the MEC, there are no explicit standardized mechanisms to dynamically allow sharing of MEC resources among different service operators or application providers. Standardizing the way MEC providers trade their spare capacity to various application and cloud providers will help generate new revenue streams and provide additional ways to recover the costs associated with MEC deployments and operations. This trading can only happen if there is both an efficient, safe and fair resource-sharing mechanism among multiple operators and a marketplace with well-defined APIs that allow MEC operators to dynamically auction off their excess capacity based on traffic projections. To ensure settling, an assurance framework is also needed that can measure the performance of complex service chains and service meshes so that the customer receives what was promised and the complex ecosystem of virtualized and containerized services works in harmony.

Unfortunately, as an industry, we are still struggling to address issues such as testing, analyzing and assuring NFV performance; the necessary first steps toward a functioning marketplace. While there has been some research in creating blockchain-based trading marketplaces, we still lack a true marketplace with integrated measurements, assurance, and settlements as well as an ability to isolate systems in case of faults.

Communication is becoming another essential utility, just like electricity and water, with a high 鈥渓ast-mile鈥� service cost. Therefore, it is important for us in the communication industry to start thinking in those terms. We still have a long way to go until we can plug any VNF/service into any NFV/container infrastructure, know how much it will approximately cost and how it will perform, and have mechanisms to allow providers to share and minimize unused resources thus maximizing profitability.

Reference

[i]

[ii]

[iii]

听