Programmable data plane hardware creates new opportunities for infusing intelligence into the network. This raises a fundamental question: what kinds of computation should be delegated to the network?In this paper, we discuss the opportunities and challenges for co-designing data center distributed systems with their network layer. We believe that the time has finally come for offloading part of their computation to execute in-network. However, in-network computation tasks must be judiciously crafted to match the limitations of the network machine architecture of programmable devices. With the help of our experiments on machine learning and graph analytics workloads, we identify that aggregation functions raise opportunities to exploit the limited computation power of networking hardware to lessen network congestion and improve the overall application performance. Moreover, as a proof-of-concept, we propose DAIET, a system that performs in-network data aggregation. Experimental results with an initial prototype show a large data reduction ratio (86.9%-89.3%) and a similar decrease in the workers' computation time.
Motivated by the importance of accurate identification for a range of applications, this paper compares and contrasts the effective and efficient classification of network-based applications using behavioral observations of network-traffic and those using Deep-Packet Inspection.Importantly, throughout our work we are able to make comparison with data possessing an accurate, independently-determined ground-truth that describes the actual applications causing the network-traffic observed.In a unique study in both the spatial-domain: comparing across different network-locations and in the temporal-domain: comparing across a number of years of data, we illustrate the decay in classification accuracy across a range of application-classification mechanisms. Further, we document the accuracy of spatial classification without training-data possessing spatial diversity.Finally, we illustrate the classification of UDP traffic. We use the same classification approach for both stateful flows (TCP) and stateless flows based upon UDP. Importantly, we demonstrate high levels of accuracy: greater than 92% for the worst circumstance regardless of the application.
The power consumption of the Internet and datacenter networks is already significant, and threatens to shortly hit the power delivery limits while the hardware is trying to sustain ever-increasing traffic requirements. Existing energyreduction approaches in this domain advocate recomputing network configuration with each substantial change in demand. Unfortunately, computing the minimum network subset is computationally hard and does not scale. Thus, the network is forced to operate with diminished performance during the recomputation periods. In this paper, we propose REsPoNse, a framework which overcomes the optimalityscalability trade-off. The insight in REsPoNse is to identify a few energy-critical paths off-line, install them into network elements, and use a simple online element to redirect the traffic in a way that enables large parts of the network to enter a low-power state. We evaluate REsPoNse with real network data and demonstrate that it achieves the same energy savings as the existing approaches, with marginal impact on network scalability and application performance.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.