Scaling Up

---

By contrast, vertically scaled-up systems consolidate many small servers into one or more large servers. Systems that scale up are characterized by a single server with multiple CPUs running a single copy of the operating system. Although scaling out has become an accepted approach to modern computing, some consider it unsuitable for certain types of large-scale, enterprise-type environments, primarily because of the disadvantages discussed in the previous section. This is where the concept of scaling up or vertical scaling comes in. Systems that scale up are characterized by a single server with multiple CPUs running a single copy of the operating system. Scaling up decreases the elapsed time that it takes to complete high-volume tasks such as online transaction processing (OLTP) and complex tasks such as weather forecasting.

Which applications are most suited to vertical scaling? According to Aberdeen Group (Boston, MA), scaling up allows IT managers to run applications and databases that were optimized for a single environment, as is the case with many of today's most popular enterprise business applications, including Oracle, PeopleSoft, SAP, etc. These applications run best when they have access to plenty of local (nondistributed) processors as well as local data. Large databases, ERP, and other single-threaded applications in particular are ideally suited to vertically scaled systems. Within data centers, especially, scaling up is probably the best strategy.

---