Browse Prior Art Database

Client server periodical requests balanced based on the server load Disclosure Number: IPCOM000240947D
Publication Date: 2015-Mar-13
Document File: 2 page(s) / 45K

Publishing Venue

The Prior Art Database


In this article we present a solution for balancing periodical requests in a client-server architecture. A large number of clients communicate with the server in regular time intervals. Server is experiencing periods of overload followed by periods of underload caused by the uneven distribution of the requests. In our solution the server as part of the response, sends back information about a time-shift for the next communication that will be initialized by the client, which is calculated based on the historical server load.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Client server periodical requests balanced based on the server load


In today's industry seeing computer environments with large number of nodes is not rare. With the use of either a public or private Cloud managing such environments is not only possible, but also simplified by the different means of automation. In particular, Cloud makes it possible to stop or start all the nodes with a single request. As the result all the nodes might become online within a similar timespan, which will cause all the applications installed on these nodes to also start within a similar timespan.

In a client-server architecture, a typical communication scheme is that clients communicate with the server in constant and repeating intervals. If most of the clients begin the communication at similar time, as the outcome the server will become flooded

with requests at one time, followed by quiet periods with only few requests. The operating system where the server is running, might be also experiencing periodical load increases, for example during nightly backups or security scans, that are typically performed on a regular basis.

If we have a system that works under uneven load we will be either wasting resources, because we need more recourses for the times when the system is under heavy load, that will be wasted during times of underload, or we will be experiencing performance issues, like communication timeouts, when there are not enough resources to handle all the requests. In both cases the system is not optimal.


We have been working with a monitoring client-server application where the clients (agents) are responsible for collecting data and sending it to the server in 5 minutes intervals, and the server is responsible for processing and presenting that data. Because this is a monitoring solution, our server already has a lot...