Quality of service aware rate throttling of delay tolerant traffic for energy efficient routing

Abstract

The invention is directed to energy-efficient network processing of delay tolerant data packet traffic. Embodiments of the invention determine if an aggregate of time critical traffic flow rates and minimum rates for meeting QoS requirements of delay tolerant traffic flows exceeds a combined optimal rate of packet processing engines of a network processor. In the affirmative case, embodiments set the processing rate of individual packet processing engines to a minimum rate, such that the cumulative rate of the packet processing engines meets the aggregate rate, and schedule the delay tolerant flows to meet their respective minimum rates. Advantageously, by throttling the processing rate of only delay tolerant traffic, energy consumption of network processors can be reduced while at the same time QoS requirements of the delay tolerant traffic and time critical traffic can be met.

Description

RELATED APPLICATIONS[0001]This application is a continuation of application Ser. No. 12 / 794,268, filed on Jun. 4, 2010, the entire disclosure of which is hereby incorporated herein by reference for all purposes.FIELD OF THE INVENTION[0002]The invention is directed to data packet networks, particularly to processing data packets in a network in a manner that respects quality of service (QoS) commitments of certain data packet traffic, referred to herein as a QoS-aware manner, while at the same time addressing energy efficiency of network processors handling the data packet traffic.BACKGROUND OF THE INVENTION[0003]The vast majority of current Internet usage consists of content being disseminated from a source to a number of users. To meet growing demands, service providers (e.g., Google) are ushered to invest in large data centers with hundreds of thousands machines distributed across different geographic regions. Similarly, network providers have been increasing network capacity by d...

AI Technical Summary

Benefits of technology

[0009]Advantageously, unlike dynamic resource provisioning methods that simply adjust processing rates based on current traffic volume demands, by throttling the processing rate of only delay tolerant traffic to result in an energy-efficient rate of the packet processing engines that are processing the aggregate traffic, energy consumption of network processors can be reduced while at the same time QoS requirements of the delay tolerant traffic and time critical traffic can be met.

[0020]Various embodiments are described wherein: identifying the delay tolerant flow includes identifying a plurality of delay tolerant flows being processed by the network processor at a first aggregate flow rate; determining a reduced flow rate for the delay tolerant flow includes determining an aggregate reduced flow rate for the plurality of delay tolerant flows based on QoS requirements associated with the plurality of delay tolerant flows; and reducing the processing rate of the network processor based on the reduced flow rate includes reducing the processing rate of the network processor based on the aggregate reduced flow rate.

[0023]Various embodiments are described wherein the network processor includes a plurality of packet processing engines (PPEs) and setting the processing rate of the network processor to substantially equal the summed rate includes reducing a rate of at least one PPE of the plurality of PPEs based on the summed rate.

Problems solved by technology

Networking devices such as servers, core / edge routers, and access devices involved in this large scale content delivery consume significant energy; for instance, the U.S. Environmental Protection Agency (EPA) estimates that data centers could consume 100 billion kilowatt hours at a cost of $7.4 billion per year by 2011.

However, current DVS techniques do not take into account QoS commitments of the traffic being routed, which commitments if not met can affect a network providers' revenues.

Images