Browse Prior Art Database

A System for Dynamically Predicting Performance of an Enterprise Service Bus Disclosure Number: IPCOM000194529D
Publication Date: 2010-Mar-29
Document File: 2 page(s) / 45K

Publishing Venue

The Prior Art Database


Enterprise Service Bus (ESB) middleware enables applications to communicate with each other by providing services to route and transform messages, and perform other complex connectivity scenarios. However, the range of capability of the ESB is sufficiently large that it is often difficult to size the computing resources required for a new ESB deployment. Tools tend rely on past experience, but as ESB implementations are rarely the same, the results are not always accurate. This article proposes a way to harness the information from past ESB deployments to more effectively size subsequent deployments, by inspecting at a low level the technical implementation of each solution.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 56% of the total text.

Page 1 of 2

A System for Dynamically Predicting Performance of an Enterprise Service Bus

It is useful to be able to predict performance of Enterprise Service Bus (ESB) systems, for example, for capacity planning, for like-for-like performance tuning (how a single variable can affect performance of a system), or for evaluating relative performance of different ESBs for common scenarios. There are different techniques for doing this, which typically involve the use of historical information of other installations. However, these tools often rely on qualitative measurements or are based on manual input or a limited set of metrics, which means that the results are prone to error or skewing.

This article proposes two tools which work together:

The first tool is able to use the ESB's administrative interfaces to inspect the configuration of an existing ESB, specifically, the following three aspects:
1) metrics relating to the complexity of the ESB and any connectivity applications deployed to it. These include the number and type of mediation primitives, size of any transformation logic, number of operating system processes and threads associated with the ESB.
2) environmental characteristics of the ESB and the platform on which it is deployed, such as the type and version of the ESB, its operating system, the processor speed and available and used memory.
3) the achieved performance of the system, including number and size of messages processed over a given time period.

    The second tool accepts as input the information provided from the tool above, as run on multiple ESB installations. When aggregated in sufficient numbers, the combination of the three sets of data listed above allows the tool to predict the expected performance of new ESBs. The user of such a tool is able to define a subset of these properties as invariant (such as the complexity of the transformation logic, the operating system, or desired throughput rates), and based on the collected information previously accumulated, have the tool calculate the remaining aspects of the configuration.

    The information provided to the second tool is not specific to any individual ESB, which means that different ESB products could each provide statistical...