Scale-Up

Scale-up, aka scale vertically, is a term used to indicate a scaling strategy based on

increasing the computing power of a single server by adding resources in term of

more processors, memory, I/O devices, etc.

Such vertical scaling of existing systems also enables them to leverage virtualization

technology more effectively, as it provides more resources for the hosted

operating systems and applications.

Scale-up limits the number of management points, potentially easing security

policy enforcement.

While attractive for some applications, most data centers prefer to use standard

components and adopt a scale-out approach.



Scale-Out

Scale-out, aka scale horizontally, is a term used to indicate a scaling strategy based

on increasing the number of servers. It has the big advantage that the administrator

can more easily re-purpose compute as required.

Scale-out strategies are commonly adopted in association with Intel X86 servers.

These types of servers, based on the PC architecture, have seen in the recent years

a continued price drop and performance increase. These “commodity” systems

have now reached a computing capacity sufficient to run the majority of applications

present in data centers. They may also be interconnected in clusters to

perform HPC (High Performance Computing) applications in scientific fields like

modeling and simulation, oil and gas, seismic analysis, and biotechnology, previously

possible only on mainframes or supercomputers.

The scale-out model has created an increased demand for shared data storage with

very high I/O performance, especially where processing of large amounts of data

is required, such as for databases.


看完这段，我的理解是一个是横向，一个是纵向

scale up类型能更好的管理资源，旨在提高单台服务器的性能

scale out则是通过增加服务器的数量，以达到整体性能提高的目的，但是因为增加了服务器的数量，在管理方面显得不如scale up。

scale up 意味着需要更大的计算机或者更强的计算能力的计算机；

scale out意味着需要更多的计算机。