The last decade has witnessed an explosive growth in database engines optimized for main memory based execution. With the current size and cost of DRAM, many analytic and transactional data sets now fit completely in memory, resulting in a vastly different set of design tradeoffs for databases compared to their earlier disk-based counterparts. There is no disk IO to overlap computation with in these systems, and they are therefore highly sensitive to CPU performance, memory bandwidth, cache sizes, network latency, etc.
The 1st annual Accelerating In-Memory Databases (AIMD) workshop aims to bring together researchers and practitioners in the area of building novel acceleration technologies for inmemory databases, for analytics, OLTP, hybrid workloads and other emerging use cases including IoT, machine learning, geo-spatial and graph applications.
AIMD 2018 will be held April 6, 2018 in conjunction with ICDE 2018.
Relevant topics include, but are not limited to, the following:
Novel data formats and data processing algorithms optimized for modern hardwareHardware accelerators such as FPGAs, GPUs and ASICs applied to in-memory databasesNovel ideas for Non-Volatile Memory based storage managersLock-free optimistic concurrency control mechanisms, perhaps exploiting hardware mechanisms such as transactional memoryIn-Memory compression algorithms and accelerators for compression and decompressionAccelerating distributed transactions using new network technologies such as RDMA and Network embedded processorsIn Memory Analytics and Hybrid Workload query processing and optimization techniques especially involving novel applications of SIMD vector instructionsHardware and software acceleration techniques for domain specific In-Memory Databases such as Analytics, Transaction Processing, IoT, Machine Learning, Spatial and Graph applications, etc.
04月16日
2018
04月19日
2018
注册截止日期
留言