High Througput Multicore Server Dengan Kernel Bypass
DOI:
https://doi.org/10.30646/sinus.v18i2.472Keywords:
kernel bypass, hashing rx, multiple ip receiver, multiple thread receiver, multiple port listenerAbstract
Kernel performs one part of the overall operating system utility as a bridge between the user and the hardware. In multi core hardware, the kernel was also created with a general approach so it needs to be adjusted for specific purposes. Kernels for servers using common Kernels have many limitations, such as single socket descriptor, single IRQ, and lack of pooling so that they require some modification to run optimally. Kernel bypass is a method by eliminating the automation of a server used to realize a server, namely high throughput. Kernel bypass is a combination of techniques, modification at the driver level with hashing rx signal and modification of multiple receivers with multiple ip receivers, multiple thread receivers, and multiple port listeners. This combined modification makes the server more reliable with an average throughput increase of 250.44%.
References
Angelo, G. D., Marchetti-spaccamela, A., & Cnr, I. (2016). Multiprocessor Real-Time Scheduling with Hierarchical Processor Affinities. 2016 28th Euromicro Conference on Real-Time Systems, 237–247. https://doi.org/10.1109/ECRTS.2016.24
Bo, Z. (2016). Analysis of the Resource Affinity in NUMA Architecture for High Performance Network. 2016 5th International Conference on Measurement, Instrumentation and Automation, 547–550.
Diener, M., Cruz, E. H. M., Alves, M. A. Z., Navaux, P. O. A., Busse, A., & Heiss, H. U. (2016). Kernel-Based Thread and Data Mapping for Improved Memory Affinity. IEEE Transactions on Parallel and Distributed Systems, 27(9), 2653–2666. https://doi.org/10.1109/TPDS.2015.2504985
Fusco, F., & Deri, L. (2010). High Speed Network Traffic Analysis with Commodity Multi-core Systems. proceedings of the 10th ACM SIGCOMM conference on Internet measurement, 218–224.
Galagan, V., Yurchenko, O., Preobrazhensky, E., Zhuravkov, P., & Dombrougov, M. (2013). Multi-gigabit intel-based software routers. Proceedings - RoEduNet IEEE International Conference. https://doi.org/10.1109/RoEduNet.2013.6714193
Gu, Q., Wen, L., Dai, F., Gong, H., Yang, Y., Xu, X., & Feng, Z. (2014). StackPool: A high-performance scalable network architecture on multi-core servers. Proceedings - 2013 IEEE International Conference on High Performance Computing and Communications, HPCC 2013 and 2013 IEEE International Conference on Embedded and Ubiquitous Computing, EUC 2013, 17–28. https://doi.org/10.1109/HPCC.and.EUC.2013.13
Hanford, N., Ahuja, V., Farrens, M., Ghosal, D., Balman, M., Pouyoul, E., & Tierney, B. (2015). Improving network performance on multicore systems : Impact of core affinities on high throughput flows. Future Generation Computer Systems. https://doi.org/10.1016/j.future.2015.09.012
Huang, C., Yu, X., & Luo, H. (2010). Research on high-speed network data stream capture based on multi-queue NIC and multi-core processor. ICIME 2010 - 2010 2nd IEEE International Conference on Information Management and Engineering, 2, 248–251. https://doi.org/10.1109/ICIME.2010.5477440
Jie, L., Shuhui, C., & Jinshu, S. (2016). Implementation of TCP large receive offload on multi-core NPU platform. 2016 International Conference on Information and Communication Technology Convergence, ICTC 2016, 258–263. https://doi.org/10.1109/ICTC.2016.7763481
Rivera, D., Ach, E., & Bustos-jim, J. (2014). Analysis of Linux UDP Sockets Concurrent Performance. 2014 33rd International Conference of the Chilean Computer Science Society, 65–69.
Shambharkar, S. A. (2015). A Study on Setting Processor or CPU Affinity in Multi-Core Architecture for Parallel Computing. International Journal of Science and Research, 4(5), 2013–2016.
Tsai, W. Y., Huang, N. F., & Hung, H. W. (2012). A port-configuration assisted NIC IRQ affinitization scheme for multi-core packet forwarding applications. GLOBECOM - IEEE Global Telecommunications Conference, 2547–2552. https://doi.org/10.1109/GLOCOM.2012.6503500
Downloads
Published
How to Cite
Issue
Section
License
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
Â
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
