Issue link: https://www.e-digitaleditions.com/i/1414667
22 I ICT TODAY 5G ELEVATES DATA CENTER CRITICALITY 5G has completely revolutionized the way communi- cations networks are built. From the core to the 5G radio, the network has been virtualized and network functions decomposed into microservices (Figure 2). In doing so, the network becomes extremely dynamic and massively scalable, which are two key features network planners and architects believe will be essential to 5G success. A third architectural feature that fundamentally changes how these networks are built and operated is network slicing. Network slicing enables virtualized, dedicated customer or application specific networks to be rapidly created to address stringent performance and availability requirements—and 5G introduces features that enable this to be delivered on-demand and at scale. center close to the end user, while other non-real-time functions, such as authentication, may be centralized in a core network data center. As the end user moves throughout the network or as performance changes, these microservices will need to be relocated to maintain real-time performance. Additionally, because of the agile nature of service building, it is expected that the network will need to be able to massively scale up or down rapidly to address unexpected demand or large-scale events, such as the Olympics or the World Cup. Finally, 5G introduces three classifications of services: • Enhanced mobile broadband (eMBB), providing data rates of 1Gb/s or more to the end user. • Ultra-reliable low-latency communications (URLLC), providing sub-ms round trip latency and 5-9s or greater availability. • Massive machine type communications (mMTC), supporting 1 million devices per square kilometer (≈0.39 square mi). FIGURE 2: Microservice decomposition and distribution. What does this mean for the data center and DCI? In this new networking paradigm, customer or appli- cation specific services are dynamically assembled by combining microservices to build the functionality required. These microservices can be dynamically distributed between data centers within the virtual network to adapt to changing service requirements and performance. For example, in an augmented reality application, the microservices involved with real-time processing of visualization will likely be found in an edge data Each of these service types place unique strains on the DCI, and the agile nature of new service devel- opment will drive additional, unforeseen DCI require- ments, thereby requiring a flexible and adaptive monitoring system to keep pace and assure that SLAs are met.