PROJECT TITLE :

ATME: Accurate Traffic Matrix Estimation in Both Public and Private Datacenter Networks - 2018

ABSTRACT:

Understanding the pattern of finish-to-finish traffic flows in datacenter networks (DCNs) is crucial to several DCN designs and operations (e.g., traffic engineering and load balancing). However, very little research work has been done to get traffic information efficiently and nevertheless accurately. Researchers often assume the provision of traffic tracing tools (e.g., OpenFlow) when their proposals require traffic data as input, however these tools could have high monitoring overhead and consume vital switch resources whether or not they're accessible during a DCN. Although estimating the traffic matrix (TM) between origin-destination pairs using only basic switch SNMP counters could be a mature observe in IP networks, traffic flows in DCNs show totally totally different characteristics, while the massive number of redundant routes in an exceedingly DCN any complicates true. To this finish, we tend to propose to utilize resource provisioning data in public cloud datacenters and also the service placement info in non-public datacenters for deducing the correlations among high-of-rack switches, and to leverage the uneven traffic distribution in DCNs for reducing the amount of routes probably used by a flow. These permit us to develop ATME as an economical TM estimation theme that achieves high accuracy for each public and private DCNs. We tend to compare our 2 algorithms with two existing representative strategies through both experiments and simulations; the results strongly make sure the promising performance of our algorithms.