Yusuke MIYAKE / Pepabo R&D Institute, GMO Pepabo, Inc. 2019.07.25 GopherCon 2019 Optimization for number of goroutines using feedback control Principal engineer Yusuke MIYAKE @monochromegane Pepabo Proposal 4. Evaluation 5. Conclusion �4 Agenda 1. Introduction How many is the optimal number of goroutines? Performance tuning in case of pt �7 Find Goroutines Chan Chan Search Match Performance tuning in case of pt �10 • I chose stability rather than speed… How many is the optimal number of goroutines in each case? 2. Background 1. Concurrency and complexity 2. Concurrency and

HBASE-21879 Read HFile ’s Block into ByteBuffer directly. 1. Background For reducing the Java GC impact to p99/p999 RPC latency, HBase 2.x has made an offheap read and write path. The KV are allocated BucketCache firstly, if the Cache misses, go to the HFile and read the corresponding block. The workflow: reading block from cache OR sending cells to client, is basically not involved in heap memory allocations that the process of reading the Block from HFile is still copied to the heap firstly, the heap block won’t free unless the WriterThread of BucketCache flushes the Block to offheap IOEngine successfully

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin [https://en Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger community is currently Hyperledger Fabric v1.1. The benchmarking efforts that the research team did yielded a significant number of performance improvements for the Fabric v1.1.0 release that more than doubled the overall performance

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin [https://en Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger community is currently Hyperledger Fabric v1.1. The benchmarking efforts that the research team did yielded a significant number of performance improvements for the Fabric v1.1.0 release that more than doubled the overall performance

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin [https://en Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger community is currently Hyperledger Fabric v1.1. The benchmarking efforts that the research team did yielded a significant number of performance improvements for the Fabric v1.1.0 release that more than doubled the overall performance

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin [https://en Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger Fabric Performance and (such as that of the previous lifecycle model) as well as decentralized models requiring a sufficient number of organizations to agree on an endorsement policy and other details before the chaincode becomes

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin [https://en Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger Fabric Performance and (such as that of the previous lifecycle model) as well as decentralized models requiring a sufficient number of organizations to agree on an endorsement policy and other details before the chaincode becomes

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin cryptocurrency Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger community is currently Hyperledger Fabric v1.1. The benchmarking efforts that the research team did yielded a significant number of performance improvements for the Fabric v1.1.0 release that more than doubled the overall performance

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin cryptocurrency Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger community is currently Hyperledger Fabric v1.1. The benchmarking efforts that the research team did yielded a significant number of performance improvements for the Fabric v1.1.0 release that more than doubled the overall performance

validated by a consensus protocol, grouped into blocks that include a hash that bind each block to the preceding block. The first and most widely recognized application of blockchain is the Bitcoin cryptocurrency Performance of a blockchain platform can be affected by many variables such as transaction size, block size, network size, as well as limits of the hardware, etc. The Hyperledger community is currently Hyperledger Fabric v1.1. The benchmarking efforts that the research team did yielded a significant number of performance improvements for the Fabric v1.1.0 release that more than doubled the overall performance