websitedesign

Data center migration: as mobile communications and last mile bandwidth enter advanced and modern applications, people need low latency network connections, computing load is moving from centralized data center to the edge of the network. But there are a lot of misunderstandings about the edge data center. According to the analysis and exposition of the Uptime Institute chief technical officer Chris Brown, the following are several reasons for the organization to misunderstand the edge data center:

Myth 1: edge computing is a way to make the server's operating cost cheaper.
The branch pattern of the local data center does not apply to edge computing, because the edge data center is not only a local data center. The edge data center is a collection of IT assets, which is close to the end user and eventually provides services from a large data center. "
This makes the edge computing more like a distributed computing architecture. It has been seen that in the manufacturing industry, centralized computing system can process jobs without direct input, and daily operations can be directly run in the production equipment of factories. However, if the central computer fails or the connection is lost, the factory system will not be able to run for a long time. Brown says, for example, car manufacturing facilities will be idle for 24 hours because the system does not know what parts to order or what vehicles are built.
The size of the edge data centers may be small, but they are not a cheap server connected by DSL. Brown says, "the edge data center won't reduce the cost as some people think. Operators can not only use low cost, not reliable hardware, users still need to be as solid as rocks infrastructure and network. They need to be designed, built, run, and operated in a high availability mode of large area data centers. "
Edge computing helps to uninstall more delay related services away from a remote large scale data center and make it closer to the user. In a sense, they are like a content distribution network, just for the specific applications and services of the enterprise.
Misunderstanding 2: the network is not important
It is more difficult to provide a good user experience on the edge of the network than in a data center with high availability connections and power systems.
"Network connections will become more important, and the edge networks are not just wiring. In the field of data center design, people sometimes regard the network as second ideas, but for the edge data center, the enterprise needs a solid network. "Brown points out.
Building an edge network means the need to change the way the data center is managed and run. The enterprise's system is no longer in a large data center with easy access to the field operation team. It is necessary to build a more cellular network and deploy hardware in modular devices on a remote site, which takes time to complete.
Brown said, "the enterprise must consider the use of multiple network service providers and multiple connection points, each connection point can meet the needs of the business support edge data center all load, so that even if the failure or loss of a single network service provider can still provide the same high quality service. This may mean the need for a mix of wired and wireless connections to ensure access even when a route is stopped. "
A new choice of the network edge is to calculate the load and even run on the cellular base station or near the metropolitan area network, which may be the best way to provide services to users.
Misunderstanding 3: it is easy to manage the edge calculation
The edge data center is not a single pattern. An edge data center is from a single rack to twenty or thirty machines, and regardless of its size, they all need the appropriate equipment.
"Don't imagine the edge data center as a cheap, small and irrelevant thing, but we must regard every individual node as a data center. Enterprises must use commercial quality equipment for design, construction, and testing. It needs to be fully tested and put into the network to support business requirements. "Brown points out.
Hyper fusion and hybrid cloud systems such as Azure Stack are a common solution, because they solve another big problem - the management edge. Brown says, "we see more and more companies start to walk this way. With more and more of these small edge data center construction in the world, if the enterprise is not good at using the remote monitoring equipment, so the airborne analysis and automation will remind the problems of enterprises, may also provide the load transfer from one IT to another IT asset asset health based on that distribution in large enterprise management a national or global IT assets will become very difficult. "
Most data center operating modes are based on the ability of the field staff to keep the equipment shift according to the needs. This is not possible for edge computing, and enterprises can now manage many small data centers and their data center assets in various locations.
The answer is to use effective remote monitoring and a large number of automation devices. If access may become a problem, redundant hardware may also be needed.  Applications also need to provide some form of self repair, or transfer faults to nearby nodes or central data centers, so that users can maintain service access even if they are degraded due to delay time. The hybrid cloud application architecture can also provide help here, allowing the function to be migrated to a core system that is easy to manage in the event of a failure.