What are edge servers? A simple guide to the edge server, its location, and benefits

What are edge servers?
Edge servers refer to servers that run computing resources closer to users, that is, the edge of the network. 
Many different services can take advantage of this proximity, such as the Internet of Things, cloud games, CDN, RAN, and mobile core services, opening up the potential for new and innovative 5G use cases such as uRLLC.
They are traditional servers, but they run at the edge.
Instead, the central server runs in a "centralized data center". (Traditional data center)
The edge server is part of the edge cloud or MEC (Multiple Access Edge Computing). It can be a single server or a pool of servers.
Why edge servers? (benefits)
The obvious advantage of the edge server is to improve the customer experience by handling functions closer to the user, thus greatly reducing the delay.
But this is not the only benefit. Using servers closer to users can also reduce backhaul bandwidth, which is necessary in the case of a central data center. Therefore, considerable cost savings in transmission bandwidth occur.
Where is the edge server located?
The physical location of the edge servers mainly depends on the type of workload and the nature of edge computing, as described below.
The following chart from LFedge (Linux Foundation) summarizes the edge position.
Two independent networks are connected by the last mile network: "User Edge" and "Service Provider Edge".
The "user edge" goes from left to center, including all locations on the customer's premises. On the user's left border, Edge is the "device edge" (edge devices such as IoT devices), and on the right side of the user edge is the "On- prem data center edge".
The user edge is different from the service provider edge because the latter is located inside the service provider's premises, but closer to the customer's premises. The service provider owns and manages it.
Any computing server located in an edge data center is called an edge server. More specifically, they are the following three locations:
Local edge server
These servers are part of the "On-Prem edge data center". They are aggregation points within the customer's premises. An example is a private 5G network location at a customer's site, which can host different types of enterprise applications.
Access edge servers
These servers are part of the Access Edge Data Center. On the edge of the network. They are very close to customers, but within the service provider's network, so they are managed by the service provider. Examples include micro data centers hosted at or closer to cellular base stations. Typical computing workloads include RAN services, such as RU, DU, delay-sensitive games, and self-driving cars.
Regional edge server
These edge servers are part of the Regional Edge Data Center. They are aggregation points of multiple Access Edge servers. The services they can run are not as critical to latency as those running in the Access Edge data center.
All-in-one white box as edge servers
With the dispersion of computing load in the network, the number of edge data centers may increase rapidly, thus increasing the cost of any telecom company. Plus space and electricity are usually the smallest edge positions.
In addition, the limited space of the micro data center does not allow the central data center to have a private network structure composed of backbone switches and leaf switches.
It will be helpful to use an all-in-one smart white box solution for MEC sites. The economies of scale of white box are beneficial to reduce costs. However, such a solution should support powerful computing and powerful integrated network blades to support the computing and network requirements of edge data centers.