Generalized Suffix Tree based Multiple Sequence Alignment for Service Virtualization

TL;DR

This paper introduces a novel multiple sequence alignment algorithm based on Generalized Suffix Trees, significantly improving speed and accuracy for service virtualization and potentially benefiting other domains.

Contribution

The paper presents a new GST-based MSA algorithm that outperforms standard methods in speed and accuracy for service virtualization tasks.

Findings

Up to 50 times faster than standard MSA algorithms

Effective in deriving message prototypes from system interactions

Applicable to domains beyond service virtualization

Abstract

Assuring quality of contemporary software systems is a very challenging task due to the often large complexity of the deployment environments in which they will operate. Service virtualization is an approach to this challenge where services within the deployment environment are emulated by synthesising service response messages from models or by recording and then replaying service interaction messages with the system. Record-and-replay techniques require an approach where (i) message prototypes can be derived from recorded system interactions (i.e. request-response sequences), (ii) a scheme to match incoming request messages against message prototypes, and (iii) the synthesis of response messages based on similarities between incoming messages and the recorded system interactions. Previous approaches in service virtualization have required a multiple sequence alignment (MSA) algorithm as a means of finding common patterns of similarities and differences between messages required by all three steps. In this paper, we present a novel MSA algorithm based on Generalized Suffix Trees (GSTs). We evaluated the accuracy and efficiency of the proposed algorithm against six enterprise service message trace datasets, with the proposed algorithm performing up to 50 times faster than standard MSA approaches. Furthermore, the algorithm has applicability to other domains beyond service virtualization.