image courtesy of openedgecomputing.org
This article is intended to be a simple introduction to what can be a complicated technical process. It usually helps to begin the articles I write involving a specific cloud technology with a definition. Edge computing’s definition, like many other technologies, has and evolved in a very short period of time. In the past Edge Computing could describe devices that connect to the local area network (LAN) to the wide area network (WAN), such as firewalls and routers. Today’s definitions of Edge Computing are more focused on the cloud and how to overcome some of the challenges of cloud computing. The definition I will use as a basis for this article is to bring computer and data storage resources as close as possible to the people and machines that require the data. Many times, this will include creating a hybrid environment for a distant or relatively slow public cloud solution. The hybrid environment will consist of an alternate location with resources that can provide the faster response required.
Benefits of Edge Computing
The primary benefit of living on the edge is increased performance. This is most often defined in the networking world as reduced latency. Latency is the time it takes for data packets to be stored or retrieved. With the growth of Machine Learning (ML), Machine to Machine Communications (M2M) and Artificial Intelligence (AI), exploding latency awareness has increased across the industry. A human working at a workstation can easily tolerate a data latency of 100-200 milliseconds (MS) without much frustration. If you’re a gamer you would like to see latency at 30 MS or less. Many machines and the applications they run are far less tolerant to data latency. The latency tolerance for machine-based applications can range from tolerating 10 MS to no latency, needing the data in real time. There are applications humans interface with that are more latency sensitive, a primary example being voice communications. In the past decade business’ demand for Voice Over Internet Protocol (VOIP) phone systems has grown which has in turn driven the need for better managed low latency networks. Although data transmission speeds are moving at the speed of light, distance still matters. As a result, we look to reduce latency for our applications by moving the data closer to the edge and its users. This can then produce the secondary benefit of reduced cost. The closer the data is to the applications the fewer network resources required to transmit the data.
Use Cases for Edge Computing
Content Delivery Networks (CDN) are thought of as the predecessor of the edge computing solutions of today. CDN’s are a geographically distributed network of content servers designed to deliver video or other web content to the end user. Edge Computing seeks to take this to the next step by delivering all types of data in even closer to real time.
The Internet of Things (IOT) devices is a large part of what is driving the demand for Edge Computing. A common application involves a video surveillance system an organization would use for security. A large amount of data is stored in which only a fraction is needed or will be accessed. An edge device or system collects all the data, stores it, and only transfers the data needed to a public cloud for authorized access.
Cellular networks and cell towers provide another use case for Edge Computing. Data for analysis are sent from subscriber phones to an edge system at the cell site. Some of this data is used immediately for call control and call processing. Most of the data, which is not time sensitive are then transmitted to the cloud for analysis later.
As the technology and acceptance for driverless cars increase a similar type of edge strategy will be used. However, with driverless applications, the edge devices will be located in the car because of the need for real time responses.
These examples all demonstrate the need for speed is constantly increasing and will continue to grow in our data applications. As fast as our networks become there will always be the need to hasten processing time for our critical applications.
If you would like to talk more about strategies for cloud migration contact us at:
Jim Conwell (513) 227-4131 firstname.lastname@example.org
we listen first…