Interaction Traces Mining for Efficient System Responses Generation

Abstract

Software service emulation is an emerging technique for creating realistic executable models of server-side behaviour. It is particularly useful in quality assurance and DevOps, replicating production-like conditions for large-scale enterprise software systems. Existing approaches can automatically build client-server and server-server interaction models of complex software systems directly from analysis of service interaction trace data. However, when these interaction traces become large, searching an entire trace library to generate a run-time responses can become very slow. In this paper we describe a new technique that utilises data mining, specifically clustering algorithms, to pre-process large amounts of recorded interaction trace data. With the obtained clusters we facilitate efficient yet well-formed runtime response generation in our Enterprise System emulation environment. We evaluate our approach using two common application-layer protocols: LDAP and SOAP. Our experimental results show that by utilising clustering techniques in the pre-processing step, the response generation time can be reduced by 99% on average compared with existing approaches.

A positional keyword-based approach to inferring fine-grained message formats

“…Service virtualization is a critical technology for DevOps and Continuous Delivery [1], to enable au-dependent systems. Service virtualization [2,3,4] requires the understanding of message formats in order to decode request messages and formulate appropriate response messages. In the domain of network security [5,6,7], intrusion detection systems (IDS) and firewall systems require the knowledge of protocol message formats before performing deep packet inspection.…”

“…P-token leverages the positions of keywords in the message template structure to differentiate a keyword occurrence from its string value's occurrences in message payload and therefore identify message keywords more accurately. P-token has three major steps with corresponding techniques: (1) identifying positional keywords, i.e., keywords together with their positions in the messages, (2) clustering the messages into groups with a two-level hierarchical clustering strategy, such that each group has high homogeneity (representing a particular type of messages), and (3) inferring the message format for each cluster based on the natural positions of keywords in messages. In Step (1), we introduce a new technique called positional token.…”

Generalized Suffix Tree Based Multiple Sequence Alignment for Service Virtualization

“…We have been working on an automated approach to service emulation which uses no explicit knowledge of the services, their message protocols and structures, but solely relies on recordings of interactions between a system under test and its environment services [3], [4]. One key aspect of the approach is to identify similarities between messages and exploit these similarities in response generation.…”

“…One key aspect of the approach is to identify similarities between messages and exploit these similarities in response generation. Commonly used multiple sequence alignment (MSA) techniques from biology [5], [6] were used to do so -refer to [4] for the details of the approach.…”

“…In order to improve efficiency, we extended to original approach by clustering the trace recordings into groups of similar messages (ideally of the same operation type) and then formulate a single representation for the request messages for each cluster [4]. This accelerated run-time performance by enabling incoming requests from the system under test to be compared only to the cluster representations, rather than the entire transaction library.…”