{"title":"Implementation of RAM-based Cache at Write-back Mode Using Virtual-NAS for DAS-based Storage on VMware Platform","authors":"Marcel","doi":"10.1109/ISRITI48646.2019.9034613","DOIUrl":null,"url":null,"abstract":"Storage was one of the crucial components that have a big impact on overall system performance, especially in a virtualization environment. The application of cache as storage performance accelerator was an option especially for small scale environments with limited resources and the usage of DAS-based storage scenarios (Direct-Attached Storage). Maximum performance obtained when the cache function was applied to both read and write operations (at write-back mode), but there was a limitation for RAM-based cache implementation on the DAS-based storage scenario with VMware-based platform which currently only supports write- through mode (acceleration was only for read operations). This paper tried to propose an alternative solution using a virtual NAS approach to apply RAM-based cache that runs in writeback mode at VMware platform on DAS-based storage scenario. Performance test performed using the workload simulation tool for single-cache and multi-cache implementation scenarios. The test results showed a significant performance improvement for read and write operations compared to baseline (without cache condition) within the scope of the workload simulation being performed. A single cache implementation scenario, the improvement range for read operation was between 27.71x - 98.02x, while write operation was between 21.94x - 63.21x better than baseline. In a multi-cache implementation scenario, the range of performance improvement for reading operations was between 110.37x - 431.43x, whereas for write operations it was in the range of 23.14x - 391.76x better than baseline.","PeriodicalId":367363,"journal":{"name":"2019 International Seminar on Research of Information Technology and Intelligent Systems (ISRITI)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 International Seminar on Research of Information Technology and Intelligent Systems (ISRITI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISRITI48646.2019.9034613","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Storage was one of the crucial components that have a big impact on overall system performance, especially in a virtualization environment. The application of cache as storage performance accelerator was an option especially for small scale environments with limited resources and the usage of DAS-based storage scenarios (Direct-Attached Storage). Maximum performance obtained when the cache function was applied to both read and write operations (at write-back mode), but there was a limitation for RAM-based cache implementation on the DAS-based storage scenario with VMware-based platform which currently only supports write- through mode (acceleration was only for read operations). This paper tried to propose an alternative solution using a virtual NAS approach to apply RAM-based cache that runs in writeback mode at VMware platform on DAS-based storage scenario. Performance test performed using the workload simulation tool for single-cache and multi-cache implementation scenarios. The test results showed a significant performance improvement for read and write operations compared to baseline (without cache condition) within the scope of the workload simulation being performed. A single cache implementation scenario, the improvement range for read operation was between 27.71x - 98.02x, while write operation was between 21.94x - 63.21x better than baseline. In a multi-cache implementation scenario, the range of performance improvement for reading operations was between 110.37x - 431.43x, whereas for write operations it was in the range of 23.14x - 391.76x better than baseline.