Masakuni Agetsuma

Flow Control for Higher Resiliency in Wide Area Distributed Cloud Computing

Wide area distributed cloud computing, where data are processed and stored at micro datacenters (DCs) located around terminals, improves response time and reduces the amount of traffic in wide area networks. However, it is difficult for the cloud to maintain service in the case of micro DC failure because the robustness of micro DCs is substantially lower than that of conventional DCs. We propose a flow control system that transfers the flow to other micro DCs by converting the L2, L3, and L4 headers of received packets via OpenFlow when micro DC failures occur. The proposed system enables flow to be transferred to other micro DCs even from M2M terminals such as sensors that are unable to change the destination IP address. We also developed a distributed control system to solve the critical problem of switching delay caused by the packet-in mechanism used when applying the OpenFlow to the distributed cloud. Emulation results showed that throughput in the distributed control became stable 12 times as quickly as that in the centralized control when flow was transferred.

A Method for Eliminating Metadata Cache Deallocation Latency in Enterprise File Servers

This paper presents an analysis of a performance bottleneck in enterprise file servers using Linux and proposes a modification to this operation system for avoiding the bottleneck. The analysis shows that metadata cache deallocation of current Linux causes large latency in file-request processing when the operational throughput of a file server becomes large. To eliminate the latency caused by metadata cache deallocation, a new method, called “split reclaim,” which divides metadata cache deallocation from conventional cache deallocation, is proposed. It is experimentally shown that the split-reclaim method reduces the worst response time by more than 95% and achieves three times higher throughput under a metadata-intensive workload. The split-reclaim method also reduces latency caused by cache deallocation under a general file-server workload by more than 99%. These results indicate that the split-reclaim method can eliminate metadata cache deallocation latency and make possible the use of commodity servers as enterprise file servers.