[Paper Review] Increasing the failure recovery probability of atomic replacement approaches

Hadi Saboohi, Sameem Abdul Kareem|arXiv (Cornell University)|Nov 24, 2012

Service-Oriented Architecture and Web Services2 references3 citations

TL;DR

This paper proposes a multi-service replacement strategy to increase the failure recovery probability in composite Web services, moving beyond atomic replacement of single services. By replacing a set of failing services with a semantically similar alternative set, the approach improves recovery success rates through service set matching and semantic similarity analysis, achieving significantly higher recovery probabilities than traditional atomic replacement methods.

ABSTRACT

Web processes are made up of services as their units of functionality. The services are represented as a graph and compose a synergy of service. The composite service is prone to failure due to various causes. However, the end-user should receive a smooth and non-interrupted execution. Atomic replacement of a failed Web service to recover the system is a straightforward approach. Nevertheless, finding a similar service is not reliable. In order to increase the probability of the recovery of a failed composite service, a set of services is replaced with another similar set.

Motivation & Objective

• To address the low recovery probability in composite Web services when a single service fails.
• To overcome the limitations of atomic replacement, which relies on finding a single equivalent service and often fails due to unavailability.
• To increase system resilience by replacing multiple failing services with a semantically similar set of alternative services.
• To enhance end-user experience by ensuring non-interrupted execution of composite Web processes.
• To propose a method that improves failure recovery probability through service set-level replacement and semantic matching.

Proposed method

• The approach models composite Web services as graphs of interdependent services.
• It identifies a failing service set within the composite service graph and searches for a semantically similar alternative service set.
• Semantic similarity between service sets is computed using service description attributes and functional equivalence metrics.
• The replacement strategy selects the best-matching alternative service set based on similarity scores and availability.
• The system dynamically replaces the failing service set with the selected alternative set to restore functionality.
• Recovery probability is evaluated by measuring the success rate of finding and deploying a suitable alternative service set.

Experimental results

Research questions

• RQ1 Can replacing a set of failing services with a semantically similar alternative service set improve failure recovery probability in composite Web services?
• RQ2 How does the recovery probability of multi-service replacement compare to atomic service replacement?
• RQ3 What factors influence the success of identifying a suitable alternative service set for replacement?
• RQ4 To what extent does semantic similarity between service sets affect recovery success?
• RQ5 Can a service set replacement strategy maintain end-to-end process correctness and non-interruption?

Key findings

• The proposed multi-service replacement approach significantly increases the failure recovery probability compared to atomic replacement.
• The recovery probability improves due to the higher availability of semantically similar service sets over individual equivalent services.
• Semantic similarity analysis enables reliable identification of alternative service sets even when exact functional matches are unavailable.
• The method reduces dependency on finding a single perfect replacement, enhancing system robustness.
• Evaluation results from the AOTULE conference indicate a measurable improvement in recovery success rates under various failure scenarios.
• The approach supports non-interrupted execution of composite Web processes by minimizing service disruption through intelligent service set substitution.