Will Edge Computing Kill the Cloud?

Feeling edgy? So are some cloud providers, as they watch edge data centers cut into their market.

Edge data centers offer many advantages, including reduced latency, increased bandwidth, and enhanced privacy, security, and resiliency. But does this mean that cloud computing is going to be edged out of business? Not at all！Edge computing is not a threat to cloud providers, it’s an opportunity.”

Grace Lewis, a principal researcher at the Carnegie Mellon University Software Engineering Institute, says she doesn't view edge computing as a threat to cloud providers. Lewis sees edge computing as an opportunity to support a larger set of use cases in which computation and data are distributed across cloud servers and edge devices as needed. “It provides the opportunity for edge devices to perform localized processing and only interact with cloud services when greater computing resources are needed for sending data to the cloud, or for receiving data from the cloud,” she notes.

Edge Advantages
Organizations leveraging IoT devices, or sensors collecting large amounts of data at the edge, will be able to use their edge devices to perform localized data processing to achieve faster response times, Lewis says. “Other organizations that will greatly benefit from edge computing are military and first responder organizations in which access to the cloud cannot always be guaranteed,” she says. “In these cases, edge devices can be pre-loaded with computation and data that are needed for each mission, with reach-back to the cloud if and when available.”

Any organization that has use cases that require low latency -- such as Industrial IoT or cognitive applications -- should be exploring the use of edge computing to augment their computing architecture, Miller advised. Applications leveraging 5G technology are also well-suited to edge computing, he notes.

Miller predicts there will continue to be an explosion of edge computing use cases as application architects begin to recognize the advantage of reduced latency in cognitive and industrial IoT applications. “Develop once/deploy everywhere container-based applications will continue to simplify the ability to capitalize on multi-tiered architecture,” he says.

Cloud Advantages
Cloud providers will become more important in the years ahead, not less, predicts Michael McCarthy, assistant professor of information systems at Carnegie Mellon University's Heinz College of Information Systems and Public Policy. “Code will need to be redeployed to the edge for updates and security patches while edge components will need to talk to cloud components for applications that simply cannot run on the edge alone.”

Computers in the cloud will always be faster and far more capable than those on the edge, McCarthy says. They will [also] have access to larger data stores and faster networks. In contrast, computers on the edge will always be resource-poor in comparison to those in the cloud. Both technologies will improve over time, he says, but the cloud will always be in the lead.

Compared to edge computing, the cloud is very elastic in terms of both compute and long-term storage applications and built to handle surges easily, says Ed Fox, CTO of IT telecom company MetTel. While the edge doesn't provide the same degree of elasticity as the cloud, the technology can also add to an organization’s maintenance and operations load. “It takes additional work to keep these edges operating efficiently and build high availability and business continuity scenarios.” Other key edge drawbacks, according to Fox, are added network administration and data governance complexity.

Peaceful Co-Existence
Cloud computing and edge computing can work together. Systems will be designed so that each type of player does what it does well.For example, may be deployed to provide nearby cloud services, bringing the cloud closer to devices and end users. Cloudlets provide powerful computing along with low latency.

Lewis predicts that more cloud service providers will soon begin supporting what's being described as the cloud-to-edge continuum, in which computation and data flow from the cloud to the edge as needed to provide better support to edge users in areas such as latency, bandwidth and resiliency. “In these cases, decisions on what computation and data to push to edge devices, and when to do so, will require planning and the development of potentially complex algorithms to ensure that edge users have the computation and data they need when they need it,” she explains.

Edge computing has been embraced by all the major cloud providers, Miller says. “Each one has recognized and incorporated edge computing into their commercial and consumer product offerings.”
Fox agrees. “All the cloud providers are hyper-focused on extending their services to the edge, so it is really a growth area for them as well.”