5G, Data Center Evolution, and Life on the Edge
The initial benefit of 5G wireless technology comes from extremely fast Internet - otherwise known as enhanced mobile broadband (eMBB). Already, 5G speeds in excess of 10x over today’s current LTE networks have been demonstrated. In fact, Tier 1 wireless providers have launched 5G services having speeds of 200Mb to 1Gb+ in select cities and this continues to ramp up.
However, there is much more to 5G than speed. 5G brings new advancements for mission-critical communications and IoT applications far beyond what we have seen in the past generations of cellular communications. It is anticipated that billions of devices that directly impact our lives in terms of utility, transportation, health, and entertainment will be needed. And these devices will be attached to the network in a more synchronized, correlated, monitored manner in near real-time using artificial intelligence and machine learning in an effort to make us even more informed, productive and safer than possibly ever conceived.
5G roadmaps include improved reliability, robust scaling, and network slicing - an ability to create 1000s of virtual network instances on one common network to maximize resource utilization and support various service levels for different levels of users. Users will range from the daily consumer to SMBs (Small Medium Businesses) to large enterprises, and even governments. Perhaps the most valuable dimension of 5G will be the ability to provide ultra-reliable low latency (URLLC) networking from human to machine and machine to machine (M2M).
5G will enable new applications requiring low latency and high reliability. Some 5G services will require latency as low as 1 millisecond as shown in the figure, which is an incredible reduction (~100x) dictating the need to introduce flexibility into the RAN architecture and where the functions are placed. By placing RAN functions and applications closer to the end users, at an edge data center for example, distance and therefore latency is reduced.
Latency is simply defined by the time it takes for data requests and responses to travel through the network. Measured in milliseconds, the lower the rate of latency, or delay in response time, the better. This means an end user has a faster and more accurate real-time interactive experience on their PC or mobile phone for applications such as clicking the purchase button on an ecommerce site, driving with GPS, and multi-user online gaming. For enterprises and utilities this includes controlling factory machinery, collecting network performance analytics, and more.
The 4th Industrial Revolution will be underpinned by 5G’s enhanced and highly reliable networking features. It will expand conveniences, increase productivity, and expand businesses in a way that will boost GDPs around the world.
This next phase of wireless technology will directly and indirectly enhance all of our lives and drastically change the way we live over the next decades to come. We look to 5G to allow us the possibility of a virtually seamless interaction with each other and the things around us every day for the better.
3G and 4G completely untethered us from our homes or offices for communications, entertainment and fundamental business transactions. 5G will further enable driverless vehicles for individuals, mass transit and freight transportation. Robotics will become even more efficient and precise for manufacturing, healthcare, and emergency response. Entertainment and shopping experiences will also become far more immersive with the use of virtual and augmented reality applications. More data will become stored and available than ever.
It takes more than just a “5G” New Radio (5G NR) at the cell tower to deliver true 5G benefits as previously described. Just as 5G provides a dramatic change to how we will live in the forthcoming future - it will also require a drastic change in infrastructure and network architecture for the innovative service providers that deliver it successfully.
It requires disruption of the typical end to end elements and architecture that support today’s 3G and 4G wireless communications networks. Several challenges that are paramount to the delivery of the full menu of 5G services include spectrum availability, number of cell sites, fiber optic interconnectivity, radio cell coordination, and proximity of cloud data centers. There is a need for many more cell sites throughout our cities, both macro and small cells, and data centers – distributed closer to the edge of the network and using a smaller footprint to save on site costs. To connect these sites, service providers must also improve connectivity – converting all cell sites to high speed fiber optic cable and associated transport protocols, and upgrade to power at more locations to provide the needed capacity and coverage.
Centralized radio access networks and Mobile edge computing
Recently, I had the opportunity to participate within a panel of distinguished data center business professionals and experts at DICE West in San Jose, CA where we discussed “The Last Mile”, centralized RAN and how data centers will need to evolve to multi-access edge computing (MEC) supporting 5G technology and deployments. As part of our discussion we addressed the proximity of data centers to existing and future cell sites, and the evolution of physical network functions, like radio controllers transitioning over time to the cloud as virtual functions.
As shown in the figure, CSPs are transitioning more sites to a centralized RAN (C-RAN) architecture. According to Rethink Research, deployments of C-RAN sites outpace distributed RAN (D-RAN) sites by 2022. In doing so, operators benefit from:
- Faster deployments and cell densification: Plug-and-play radio deployments can be achieved because the data center (or C-RAN hub) can have baseband units pre-installed along with high speed routers and optical fronthaul equipment pre-previsioned for seamless cutovers of greenfield and brownfield macro and small cells.
- Improved network performance and reliability: Faster speeds and better reliability can be achieved through centralized baseband (eNodeB/gNodeB) coordination, and larger 10G to 100G+ transport connections shared amongst clusters of C-RAN cells. It also provides an outstanding option of expanding spectrum and adding radio at the cell sites without a signicant amount of construction, a key characteristic as more spectrum is released and carrier partnerships evolve to become more competitive and reach more people and devices.
- Lower installation and long-term operating expenses (OPEX): Construction and installation, and leasing costs are cut by shrinking the required footprint and power of leased cell site space during addition of 5G sites by moving baseband processing from the base of the tower to the edge data center.
- New partnerships and revenue streams: The centralized architecture is also an essential means to embrace the full benefits of cloud computing within cellular network operations and not just user-facing over-the-top applications. Using the 5G network to natively host on demand real-time applications will translate directly into new revenue generating opportunities from native CSP, enterprise and hybrid cloud hosted applications.
Since distance is one of the greatest contributors to latency to achieve URLLC, data centers need to be located closer to the cell sites so that computing can happen closer to and faster for end users. Because of this, distance to the data center’s 4G and especially 5G wireless core function has become more important, driving new roles and terms, such as “edge” and “far edge” data centers. We are already seeing operators using MEC to move applications and processing closer to the end users to reduce latency.
Instead of sending the data to a centralized data center and/or cloud for processing, the network edge analyzes, processes, and stores the data based on content and service level requirements by the end users. This push to minimize latency supports time-critical applications like autonomous vehicles, health services, and virtual and augmented entertainment.
Demand for full 5G level enhanced features such as URLLC will really drive growth in edge data centers. In the next few years, we will likely see a mix of “far edge,” “edge,” and “core” data centers located in many areas to support the diverse mix of 5G services.
AI and 5G at the edge
Our DICE West panel also discussed how artificial intelligence (AI) will impact 5G throughout the network and particularly the edge. AI can be applied throughout the end-to-end network, tailored for a specific application, service, or network operation. By combining the benefits of AI with edge computing power, service providers will be able to improve network operations and enable new service opportunities in real-time and even avoid outages and service degradation pre-emptively based upon trending network data and historical data.
AI can deliver better network management services for further enhanced security, Last Mile data transport, and core traffic routing for higher availability and lower latency user applications. AI will also assist in the speediness of network slice provisioning and respond to network outages much faster than direct human interactions.
5G operators will ultimately be able to support an end-to-end automated platform driving innovation and customizable, software-driven open edge platforms and applications that continuously provide high QoS to consumers, enterprises, and other verticals despite network congestion and conditions thus fortifying the quality of life for individuals and adding significant value to businesses.
Share your thoughts on this topic by joining the Twitter discussion with @nokianetworks or @nokia using #5G #IoT #virtual #network #industry40 #edge #DataCenter #AI #teamnokia #wirelesstechnology #enterprise #cloudcomputing #robotics