In Cloud Computing or Content Distributed Cloud (CDN) workload happens between a local web server and proxy servers. The typical method is DNS redirecting and the workload factoring decision is predefined manually over a set of most popular objects. For such popular object detection in requests tails, many schemes have proposed for fast request rate estimation in traffic monitoring and fast counting for most popular data item for which request arrived and above mentioned jobs done. Accurate request rate estimation is necessary for resource planning & management, measuring compliance to SLAs and cloud security especially in Cloud Computing where we are talking about Integrated, Heterogeneous or Hybrid Cloud deployment model. With following ideals, we propose SRATE (State-Rank bAsed Traffic Estimation) has sufficient short estimation times with provable bounds on estimation error, low memory usage and also easily implementable in hardware for operation at high speeds. With developing such scheme, we achieve up to three orders of magnitude speedup in estimation time. The speedups are achieved with low memory usage by using “State-Rank (0,1,?)” instead of runs or coincidences. This new scheme has many benefits at high oppressive workload including quicker detection of unhandled oppressive data items or spikes with incipient denial of service attacks. As result, we show that the proposed scheme is faster and more accurate than other schemes. We also prove bounds on the scheme’s accuracy, request size or estimation time, average request rate estimation, memory needs, average request sending rate and also show that it performs well by efficient simulation techniques.

Content Distributed Cloud SRATE Memory Cost