Abstract
In this paper, the existing approaches to the assessment of computer networks performance are considered. The standard structure of a network of the application layer of the OSI model using the example of SBIS3 application (product of Tensor Company) is treated. Further, two approaches allowing to analyze degradations in a network are considered - on the basis of aggregated data and the operational analysis. The degradation study of more than 60 000 request types between two versions of application which works on the basis of the computer network is the cornerstone of the first decision. Each type of requests is described by four based metrics, each metrics representing a time series. The input data are aggregated every 10 minutes before an analysis algorithm. Further, the threshold criteria based on mathematical expectation and dispersion within two adjacent versions of the software are used. Such an approach allows to significantly reduce time for the analysis of potential problems in case of updates within the computer network. The second decision is based on not aggregated input data. It consists of detail information about all requests, there are data section of the computer network. A threshold criterion is based on durations in the selected queue. This analysis type allows to diagnose the errors with problem clients.
Highlights
The paper studies the problem of continuous delivery
The standard structure of a network of the application layer of the OSI model using the example of SBIS3 application is treated
Version2 was the first attempt without the Service Level Agreement (SLA) database and the percentile division
Summary
The paper studies the problem of continuous delivery. It shows how we promote our software into production more and more effectively. Every version release candidate is treated as a real productive version. Our understanding of how the Web works has led us to develop better ways of having machine-to-machine communication. Virtualization platforms allowed us to provide and resize our machines at will, with infrastructure automation giving us a way to handle these machines at scale. The main solution is setting Service Level Agreement (SLA) for testing the system [2]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
