Architecture and Infrastructure« Go back
Application and system architecture are documented using symbolic models that hide complexity and expose detail at varying levels of abstraction. Levels of abstraction typically begin as pictorial representations of systems and their relationships to other systems and infrastructure and support drilling down into implementation detail.
Different infrastructure components have their own pictorial notations. Applications have artifacts such as object diagrams and interaction diagrams based on UML standards. Databases have logical and physical models, and networks have models of their topology.
Models are developed during analysis and design and are the basis for implementation. Today's tools allow a substantial about of implementation detail to be generated from the models themselves speeding delivery and reducing errors.
As systems age but retain business value they may be migrated to newer technology platforms. Applications based on structured languages and flat files are re-architected into object oriented languages and relational databases. Proprietary communication networks and protocols may be replaced by LANs and WANs supported by industry standards.
System architecture can have substantial impact on physical plant architecture and design. Today, as data centers strain against power and cooling constraints, system architectures using server and storage virtualization plan critical roles in conserving floor space, cooling and power requirements as well as supporting greater flexibility, availability, and disaster recovery capabilities.