Digital Library

cab1

 
Title:      EVALUATION OF REST INTERFACES FOR RELIABLE INVOCATION SEMANTICS
Author(s):      Ricardo Filipe, Filipe Araujo
ISBN:      978-989-8533-57-9
Editors:      Pedro IsaĆ­as
Year:      2016
Edition:      Single
Keywords:      REST, Reliability, Invocation semantics
Type:      Short Paper
First Page:      253
Last Page:      256
Language:      English
Cover:      cover          
Full Contents:      click to dowload Download
Paper Abstract:      Nowadays, the REpresentational State Transfer (REST) architectural style plays a central role in distributed software. All major web and cloud providers, such as Facebook, Dropbox, Google, Instagram, Amazon, among many others, expose public REST interfaces for web, mobile and stand-alone clients. This raises the question of knowing the reliability of these REST services, as they run over the network, thus being inevitably prone to failures. While some of these failures might be harmless, and others might be overcome with a simple repetition of the same request, a few others, e.g., payments, need a much more careful treatment. In these cases, repeating the same request might not be an option, because duplicate executions might be worse than no execution at all. To answer this question, and determine the reliability of top-tier REST interfaces, we evaluated three real services, from leading web and cloud providers. We repeatedly invoked their REST functions for several days, to count the number of unsuccessful interactions. Despite running smoothly for most of the time, our initial results show that all these services suffered perturbations during the course of our experiment. This observation supports a straightforward conclusion: if even at this level we observe this number of problems, then, developers must definitely provide additional logic on the client and server, to ensure reliable interactions in the presence of faults. This experiment is part of a larger effort, to provide clear guidelines for reliable invocation semantics in web pages.
   

Social Media Links

Search

Login