Edge Computing Data Centers: Cooling the Future of Communications
Edge computing, or Mobile Edge Computing (MEC), is an alternative to traditional cloud computing. Rather than handling the bulk of the data processing at a central location, MEC moves that data processing closer to the consumer, or to the “edge” of the network.1 The ultimate goal of this shift is not necessarily to replace cloud computing entirely; more accurately, the goal is to improve the idea of cloud computing by merging its technology with the geographic distribution of processing centers in order to solve many of traditional cloud computing’s limitations.2
One notable limitation that MEC attempts to solve is the issue of latency. By shifting to more localized systems, data processing is performed closer to the user, thus cutting down on upload and download times and improving bandwidth performance.2 The goal is to create a more instantaneous experience for both commercial and private users; in turn, an increasingly digitally dependent global society can function more efficiently with this cost-effective, time-saving method of data storage and processing.
Network strain is also improved with a shift to MEC. Centralized servers are still necessary under this setup, but their burden is greatly reduced by the introduction of edge computing centers and a shift to individual devices (e.g., laptops, cellphones, tablets) handling the data processing.1 MEC can be strategically placed at locations where multiple networks intersect, effectively turning them into traffic hubs à la train stations or airports and creating a ring of localized data processing centers.3 The ease of localized deployment is aided by the fact that edge data centers are relatively small—about the size of a briefcase.3 Potential locations include sites that are found within the outer reaches of a network; factory buildings, hospitals, homes, and even vehicles can house edge computing hubs for otherwise hard-to-reach locales. Access requests from consumer devices go through these hubs where partial or whole information is readily processed, lessening the workloads of previously overburdened centralized servers.
In short, MEC allows for a more efficient and intuitive flow of data traffic. This will prove necessary with the growth of 5G capabilities as MEC is the key component behind the success of 5G technology.3 As 5G aims to increase processing speeds by up to 100 times that of the outgoing 4G networks,4 a shift to bring the processing power closer to the consumer is essential. This shift, and the tandem of 5G and edge computing technologies, will enable real-time data processing.5 And that shift will arrive with rapidity. Experts estimate that the switch to 5G networks will drive MEC to account for 75% of all enterprise-generated data processing by the year 2025; in 2018, MEC accounted for just 10%.6
Predictably, the speed at which 5G and MEC are taking over will put a temporary strain on existing infrastructure. As network providers make a push for 5G coverage, their expansion of that coverage is limited, in part, by the distribution of new MEC centers, and the limited capabilities of the older data processing facilities. New MEC centers are essential in not just the switch, but the transition to 5G, as the combination of 4G and MEC is seen by network providers as a steppingstone between 4G and 5G performance.7 This further emphasizes the necessity for capable MEC hardware in remote locations. And on the topic of remoteness, setting up these MEC centers creates an added headache for network providers as they look to expand 5G and its related services to more rural areas. MEC centers must be reliable enough to survive the intervals where technicians will be unable to service them, an issue that is prevalent in rural areas where a lack of infrastructure has already resulted in the so-called “rural broadband gap.”3 If 5G and MEC are to become the standards of the industry, their success in these more challenging locations will be critical.
The dependability of MEC centers will be important in other evolving technologies as well. For instance, the Internet of Things, or IoT, is a concept in which the effective implementation of MEC centers will be vital to its success. IoT describes the connection between objects containing sensors that communicate with one another, and the shift of moving processing power closer to these individual sensors has several benefits, chiefly among them a reduction in latency.8 Much like the adoption of 5G capabilities, the success of IoT and MEC are linked, as the close proximity between sensors generating data and the centers processing that data allows for the successful implementation of the technologies that rely upon them; both consumer and residential technological ecosystems are able to perform more powerfully and efficiently with this type of architecture.8
MEC also plays a key role in the evolving world of military communications. As with other technological advancements, the military has remained at the forefront of innovation when it comes to its network and communication infrastructures, a necessity given the threats that they face; cyber and electronic warfare attacks can cause both short-term and catastrophic disruptions to the military’s communications.9 Countering these threats is crucial to the success of the military in the landscape of modern warfare. Leveraging MEC and IoT processes allows warfighters to improve their command post mobility by increasing processing power in remote locations and keeping them better equipped to handle incoming threats and secure their communications.9 Likewise, this shift enables the use of robotic and autonomous systems that can be sent into hazardous situations in lieu of personnel; operators are able to control these systems from a safe distance with minimal latency, an absolute necessity in mission-critical scenarios.9
In looking at future and existing data centers that could provide MEC capabilities, engineers are focused on upgrades that will ensure a high level of performance. Naturally, thermal management methods are near the top—if not at the top—of their list. These thermal management solutions must meet the power and reliability standards that MEC demands. Passive two-phase cooling is widely seen as a potential fit in this regard.10 Many MEC hardware designs call for cold plate evaporators, thermosyphons, and heat pipes to dissipate the heat generated by the powerful electronics to more active cooling approaches at the room level. This approach is seen as a cost-effective, reliable, high-performing solution.
Similarly, passive two-phase immersion cooling could be the key to unlocking MEC’s full potential, optimizing its innumerable data centers to be as energy-efficient as possible.11 This process involves submerging servers in specially designed, non-conductive fluid; as the liquid passes over the electronic components, it removes heat and evaporates, rising through the system before condensing and falling back down over the components.12 This creates a passive, closed-loop system that reduces energy costs by an average of 22.8% compared to air-cooled systems.11
With the emergence of MEC as a more efficient means of data processing and transfer, it makes sense that the cooling solutions that support its related technologies must be equally proficient in cutting down on energy costs. In comparison to in-row air cooling, passive two-phase cooling for electronics is expected to be a major improvement in terms of system and energy usage efficiencies.10 In fact, major tech companies have already begun making the switch from air cooling to passive two-phase—and in particular, immersion—cooling techniques.12 With the industry building toward a future dependent upon countless edge data centers, these energy savings will add up quickly. Two-phase cooling, then, is the key to making MEC a sustainable option for the future.
As previously mentioned, that future is the near future. Hardware companies and service providers are currently working on converting old infrastructure and developing new centers in anticipation of the widespread adoption of edge computing over the next several years. The thermal management of that infrastructure is key in its success, and the passive two-phase cooling that ACT offers would prove valuable in this emerging market.
References:
- Macri K. How Edge Computing Drives Efficiency, Enables 5G. GovernmentCIO Media and Research. 2021 August 25. Accessed August 2021. https://governmentciomedia.com/how-edge-computing-drives-efficiency-enables-5g.
- Wagner V. The Edge Ecosystem Puts Computing Where It Counts. TechNewsWorld. 2021 August 5. Accessed August 2021. https://www.technewsworld.com/story/the-edge-ecosystem-puts-computing-where-it-counts-87227.html.
- Schnitzer J. Stronger together: how the convergence of Edge data centers and 5G will revolutionize data transmission and IoT. Data Centre Dynamics. 2021 August 26. Accessed August 2021. https://www.datacenterdynamics.com/en/opinions/stronger-together-how-the-convergence-of-edge-data-centers-and-5g-will-revolutionize-data-transmission-and-iot/.
- Kudikala C. 5G vs 4G: How Is One Different From The Other?. Telecom Talk. 2021 August 20. Accessed August 2021. https://telecomtalk.info/5g-vs-4g-how-are-they-different/461122/.
- Mathew G. Edge Computing: Tech’s Next Trillion-Dollar Opportunity. Devops. 2021 August 23. Accessed August 2021. https://devops.com/edge-computing-techs-next-trillion-dollar-opportunity/.
- van der Meulen R. What Edge Computing Means for Infrastructure and Operations Leaders. Gartner Research. 2018 October 3. Accessed August 2021. https://www.gartner.com/smarterwithgartner/what-edge-computing-means-for-infrastructure-and-operations-leaders.
- Riggins J. How Edge Computing Will Deliver on the Promise of 5G. The New Stack. 2019 March 8. Accessed August 2021. https://thenewstack.io/how-edge-computing-will-deliver-on-the-promise-of-5g/.
- Casey K. Edge computing and IoT: How they fit together. The Enterprisers Project. 2021 March 8. Accessed October 2021. https://enterprisersproject.com/article/2021/3/how-edge-computing-and-iot-fit-together
- Kawasaki C. Four Future Trends in Tactical Network Modernization. S. Army. 2019 January 24. Accessed October 2021. https://www.army.mil/article/216392/four_future_trends_in_tactical_network_modernization
- Amalfi RL, Enright R, Kafantaris V. Emerging 5G Networks: Potential Economic Benefits of Two-phase Thermal Management. Electronics Cooling. 2021 April 20. Summer 2021: 19-23.
- Arroba P, Moya JM, Pérez S. Energy-conscious optimization of Edge Computing through Deep Reinforcement Learning and two-phase immersion. Future Generation Computer Systems. 2021 December. Volume 125: 891-907.
- Warren T. Microsoft is now submerging servers into liquid baths. The Verge. 2021 April 6. Accessed October 2021. https://www.theverge.com/2021/4/6/22369609/microsoft-server-cooling-liquid-immersion-cloud-racks-data-centers