2017/12/28 by Chirag Juyal, Sandeep S. Kulkarni, Juyal, Chirag +7
Computer Science · #Cloud Computing and Resource Management #Distributed #Distributed systems and fault tolerance #FOS: Computer and information sciences #Parallel #Parallel Computing and Optimization Techniques #and Cluster Computing (cs.DC)
paper · pdf · doi:10.48550/arxiv.1712.09803
openalex publication_date 2017/12/28 · openalex created_date 2022/10/05 · openalex updated_date 2026/07/28
In the modern era of multicore processors, utilizing cores is a tedious job.\nSynchronization and communication among processors involve high cost. Software\ntransaction memory systems (STMs) addresses this issues and provide better\nconcurrency in which programmer need not have to worry about consistency\nissues. Another advantage of STMs is that they facilitate compositionality of\nconcurrent programs with great ease. Different concurrent operations that need\nto be composed to form a single atomic unit is achieved by encapsulating them\nin a single transaction. In this paper, we introduce a new STM system as\nmulti-version object based STM (MVOSTM) which is the combination of both of\nthese ideas for harnessing greater concurrency in STMs. As the name suggests\nMVOSTM, works on a higher level and maintains multiple versions corresponding\nto each key. We have developed MVOSTM with the unlimited number of versions\ncorresponding to each key. In addition to that, we have developed garbage\ncollection for MVOSTM (MVOSTM-GC) to delete unwanted versions corresponding to\nthe keys to reduce traversal overhead. MVOSTM provides greater concurrency\nwhile reducing the number of aborts and it ensures compositionality by making\nthe transactions atomic. Here, we have used MVOSTM for the list and hash-table\ndata structure as list-MVOSTM and HT- MVOSTM. Experimental results of\nlist-MVOSTM outperform almost two to twenty fold speedup than existing\nstate-of-the-art list based STMs (Trans-list, Boosting-list, NOrec-list,\nlist-MVTO, and list-OSTM). HT-MVOSTM shows a significant performance gain of\nalmost two to nineteen times better than existing state-of-the-art hash-table\nbased STMs (ESTM, RWSTMs, HT-MVTO, and HT-OSTM). MVOSTM with list and\nhash-table shows the least number of aborts among all the existing STM\nalgorithms. MVOSTM satisfies correctness-criteria as opacity.\n