Background and Problem
Due in large part to the reformations introduced in HTTP/1.1 via the REpresentation State Transfer (REST) style, the World Wide Web overcame looming scalability and reliability challenges. In recent years, however, Web-based applications have evolved from distributed hypermedia systems to a rich collection of dynamic services. These new applications however have often introduced architectural dissonance – the state of a realized architecture not providing the style’s anticipated benefits. This dissonance with REST jeopardizes the reformations that REST put into place. We can partly attribute this dissonance to a lack of expressiveness and guidance from REST concerning the construction of these emerging services. Yet, if more effective and expressive guidelines coupled with accompanying frameworks and tools were at an architect’s disposal, then the impact of the dissonance could be mitigated and architectural harmony on the Web could be restored.
|
Insights and Research Questions/Hypotheses
Insight: The Web is just one in a class of systems that exhibit specific properties that allows for clear separation of control along macro and micro architectural views.
Question: Can we introduce a vocabulary to better explain the relationship between macro and micro architectures?
Hypothesis: It is possible to construct a set of additional principles that can more precisely and effectively guide the architecture in a REST-based environment.
Insight: Computation should be a first-class concept in the Web. Adding computation leads to a new set of constraints at the macro level: computational namespaces, locus of computation, and execution dynamism. It also teases at new micro constraints such as: migrating computations, constraining transformations, minimizing core functions, mitigating latency, and providing multiple interfaces.
|
Approach and Validation
Approach
Validation
|
Contributions and Schedule
Contributions
Schedule
November 2006: Topic Defense
Winter 2007: Bring Serf & Apache in line with CREST principles; Propose redesign of HTTP to better support CREST principles
September 2007: Final defense
|
Last Modified Sunday, 05-Aug-2018 13:13:16 EDT These pages were made by Justin R. Erenkrantz unless otherwise stated. This work is licensed under a Creative Commons License. These pages will look best in an XHTML 1.0 compliant browser. |
|