Data center migration,What will the future data center be?

Data center migration, industry related changes and disruptive forces, and the prediction of new data center types and specific use cases, finally put forward an effective method to prevent downtime and use innovation for future data centers.
The future data center and specific attributes that we want to accurately predict the design and operation of physical data center in the long run is obviously challenging. Since the early mainframe, the number and type of data center's form and specification have changed significantly. In the foreseeable future, we expect that the pace of this change will continue, and we will get rid of the inefficient websites of ordinary enterprises.

Some types of data centers in the coming decades and later (some of them now exist) are likely to be determined, and these data center types may dominate over the next ten years and later. Future data center types may include, but are not limited to, the following contents:
The super large scale data center (cloud service operator, but also includes other service providers).
Data centers of cloud service providers (non super large scale) and other service providers, data centers of L hosting and MTDC service providers.
Enterprise data center (special purpose, high quality website and fewer cabinet / engine room).
Edge data center (micro data center and core site).
High performance computing (HPC) and professional data center
The above data center types will be defined by the following criteria and attributes.
1. business model
The business model based on ownership will vary from data center type. For example, host hosting and service provider stations will need to provide high availability and usually require low latency services that are usually implemented by proximity and connectivity. Similarly, some enterprises have specific workload, data requirements or management issues, which indicate that they must continue to design and operate their own advanced data centers. In view of this, these enterprise organizations will then have the opportunity to innovate and customize, but these may not be applicable to business (host hosting, host hosting) operators.
2. scale
There are signs that the scale efficiency of existing large scale data center sites means that, despite the progress of computing power and the shifting of more workload to the edge, the future will continue to require ultra high speed networks. On the other hand, small micro data center sites are expected to support Internet of things and other applications more widely. However, the trend of integrating server rooms and cabinets (traditional edges) into hosting, cloud and some small data centers will continue.
3. elasticity
Elastic demand will be more closely related to business cases and functions. For example, because of the lower service level and the ability to manage availability through load balancing, a large scale data center may have a lighter physical infrastructure (reduced UPS, a lower layer of design). According to the requirements of customers, some MTDC may also establish different resilience levels in the same facility. More software based distributed flexibility will also be adopted, and at the same time, the dependence on physical infrastructure (generators, UPS) will be reduced.
4. efficiency
Efficiency will continue to be a comprehensive requirement. Some data center facilities, such as super large data center sites, will give priority to efficiency issues.  In some cases, it is higher than most other data center standards. A large scale of the large scale data center will continue to focus on sustainable development and carbon emissions by using more renewable energy sources (electricity purchase agreements, renewable energy tariffs or in some cases in the field).
5. IT density
Due to the application of AI / machine learning, high performance computing and big data, the average machine? Shelf power density (now less than 5kW) may continue to increase over time. However, some wildcard technologies, such as quantum computation, may increase computing capacity and significantly reduce power consumption requirements. Density will increasingly be associated with business functions and workload. High density areas can be created to achieve more efficient cooling and distribution. For example, HPC and other professional sites may have high-density IT devices (each rack > 25kW), and each unit space / rack consumes more energy. This means that cooling can also be tightly coupled: the demand for relatively few high-density rack.
6. geography and distribution
Many large cloud service operators have been set up in specific locations, such as Europe, or have leased data centers from MTDC providers, partly in order to comply with data regulatory regulations. For future data center types, this trend is likely to continue. Super large data center sites will also continue to build in areas where energy costs are low, tax preferences and climate conditions allow free air cooling. The edge data center capacity will be added to the centralized data center and the metropolitan area outside the core data center.