Improving SOA antipatterns detection in Service Based Systems by mining execution traces

Abstract

Abstract-Service Based Systems (SBSs), like other software systems, evolve due to changes in both user requirements and execution contexts. Continuous evolution could easily deteriorate the design and reduce the Quality of Service (QoS) of SBSs and may result in poor design solutions, commonly known as SOA antipatterns. SOA antipatterns lead to a reduced maintainability and reusability of SBSs. It is therefore important to first detect and then remove them. However, techniques for SOA antipattern detection are still in their infancy, and there are hardly any tools for their automatic detection. In this paper, we propose a new and innovative approach for SOA antipattern detection called SOMAD (Service Oriented Mining for Antipattern Detection) which is an evolution of the previously published SODA (Service Oriented Detection For Antpatterns) tool. SOMAD improves SOA antipattern detection by mining execution traces: It detects strong associations between sequences of service/method calls and further filters them using a suite of dedicated metrics. We first present the underlying association mining model and introduce the SBS-oriented rule metrics. We then describe a validating application of SOMAD to two independently developed SBSs. A comparison of our new tool with SODA reveals superiority of the former: Its precision is better by a margin ranging from 2.6% to 16.67% while the recall remains optimal at 100% and the speed is significantly reduces (2.5+ times on the same test subjects).

Detection of REST Patterns and Antipatterns: A Heuristics-Based Approach

“…A few contributions are available on the detection of SOA (anti)patterns for various SOA standards, e.g., SCA (Service Component Architecture) [3,[9][10][11] and Web services [13]. To the best of our knowledge, the detection of REST (anti)patterns, in the literature deserves yet to receive attention.…”

“…Despite the presence of several technologyspecific approaches in SCA (Service Component Architecture) and Web services (e.g., [3,[9][10][11]13]), they are not applicable for detecting (anti)patterns in REST. Indeed, the key differences between REST architecture and other SOA standards prevents the application of these approaches because: (1) traditional serviceorientation is operations-centric, whereas REST is resources-centric, (2) RESTful services are on top of JSON (or XML) over HTTP, whereas traditional Web services are on top of SOAP over HTTP or JMS (Java Message Service), (3) Web services use WSDL (Web Service Definition Language) as their formal contracts; REST has no standardised contract except the human-readable documentations, (4) traditional services are the set of self-contained software artefacts where operations are denoted using verbs; resources in REST are denoted by nouns and are directlyaccessible objects via URIs, and (5) REST clients use the standard HTTP methods to interact with resources; Web services clients implement separate client-stubs to consume services.…”

A Lightweight Approach for Specification and Detection of SOAP Anti-Patterns

“…Most state of the art techniques focused on detection of anti-patterns from object oriented software applications but these techniques are not capable of detecting anti-patterns from SOA. We identified only a few representative approaches on specification and detection of SOA anti-patterns from Web-services [9,12,13,14,15,16,18]. To the best of our knowledge, most authors used different metrics and static/ dynamic analysis methods for the detection of anti-patterns from web-services.…”

Compreensão de mudanças estruturais no código fonte usando análise dinâmica e estática

“…A manutenção de software ainda possui um alto custo e as empresas gastam muito tempo tentando entender e implementar mudanças nos sistemas (SEACORD; PLAKOSH; LEWIS, 2003; DIT; REVELLE; POSHYVANYK, 2013;MEYER, 2000). Mesmo as abordagens úteis existentes baseadas no código fonte (que detectam mudanças estruturais) e em históricos de versões (que detectam mudanças lógicas), carecem de informações comportamentais e não capturam satisfatoriamente mudanças associadas a funcionalidades específicas.…”

“…Abordagens baseadas em análise dinâmica são usadas para medir o acoplamento dinâ-mico (ARISHOLM; FOYEN, 2004), implementar novas funcionalidades (SILVA et al, 2011), tratamento de bug (ROEHM; NOSOVIC; BRUEGGE, 2015), reconstrução da arquitetura (RIVA; RODRIGUEZ, 2002) e em projetos SOA (NAYROLLES; MOHA; VALTCHEV, 2013;PAUW et al, 2005). Um algoritmo foi proposto para remover detalhes de implementação (SILVA et al, 2011).…”