在以往網路服務設備架設有著成本高昂，且常常需要專用設備，缺乏彈性，日後不論是在修改拓樸或是升級等都有著相當高的成本。所以網路功能虛擬化概念漸漸興起。使用軟體模擬這些網路服務設備的功能，讓這些網路服務功能能安裝在一般性的設備上，能夠解決以往成本高昂、彈性低等缺點外，對這些專業設備的依賴性降低也有利於更加普及。但缺點在於畢竟並非專門的硬體設計，如果效能不足以支撐網路流量，將會導致封包遺失、網路延遲增加等影響使用者體驗事情發生。所以效能方面一直都是虛擬化網路想要克服的問題。

在本研究中，想要解決的是，當網卡效能與系統處理器效能不相等時，將會產生效能瓶頸，想要去探討其中核心數量的改變對於提升效能的幫助。以及在虛擬化網路中，當流量過大，勢必需要進行分流，而使用雜湊函式的方式有可能產生相當不平衡的分配。並且想要考量其分流方式與核心數量的關聯。而本研究在網卡與核心效能不匹配的處理方式是，藉由使用多個核心進行平行處理提升整體的效能。並且在雜湊計算的部分，使用修改後的方式，藉由前一段時間的經驗也改善不平均的問題。

實驗結果也表明，當核心數量增加時，對於效能提升是有幫助的。但並非能無止境的增加，同時增加的幅度也並非相同，如何找到最符合經濟效益的數量值得討論外。在雜湊計算的部分也可以了解到，新的方式因為分配更為平均，使得每一個核心都能盡可能發揮其效能，雖然每一個核心的效能都沒提升，但因為空閒的核心減少，所以總體效能仍然是提升的。

In the past, the network service equipment was expensive and lacked of flexibility. Therefore, the concept network function virtualization is emerging. The use of software to simulate the functions of these network service devices, so that these network service functions can be installed on general equipment, can solve the shortcomings of the previous expensive and lack of flexibility. However, the disadvantage is that it is not a dedicated hardware design, so performance is a major concern compared to dedicated devices. Therefore, performance has always been a problem that virtualized networks want to overcome.

In this study, when the NIC performance is higher than the CPU performance, the packets will start to be missed. By using multiple CPUs to process network packets, the overall performance is increased. Also, the modified hash is used to balance the traffic and increase the overall performance.

The experimental results show that when the number of CPUs increases, the performance increases, but not the more CPUs, the better the modified hashing can average the traffic compared to the original one.

[1] “Fortinet,” https://www.fortinet.com/ [Accessed: July-2023].
[2] Fortinet, “Global threat landscape report. a semiannual report by fortiguard labs,”2022, https://www.fortinet.com/content/dam/maindam/PUBLIC/02 MARKETING/08Report/report-2022-H1-threat-landscape.pdf [Accessed: July-2023].
[3] U. Cisco, “Cisco annual internet report (2018–2023) white paper,” Cisco: San Jose, CA,USA, vol. 10, no. 1, pp. 1–35, 2020.
[4] F. Lau, S. Rubin, M. Smith, and L. Trajkovic, “Distributed denial of service attacks,” in Smc 2000 conference proceedings. 2000 ieee international conference on systems, man and cybernetics. ’cybernetics evolving to systems, humans, organizations, and their complex interactions’ (cat. no.0, vol. 3, 2000, pp. 2275–2280 vol.3.
[5] K. Ingham, S. Forrest et al., “A history and survey of network firewalls,” University of New Mexico, Tech. Rep, 2002.
[6] O. Sukwong, H. Kim, and J. Hoe, “Commercial antivirus software effectiveness: an empirical study,” Computer, vol. 44, no. 03, pp. 63–70, 2011.
[7] H.-J. Liao, C.-H. R. Lin, Y.-C. Lin, and K.-Y. Tung, “Intrusion detection system: A
comprehensive review,” Journal of Network and Computer Applications, vol. 36, no. 1, pp.16–24, 2013.
[8] Z. Wang and X. Li, “Intrusion prevention system design,” in Proceedings of the Interna-tional Conference on Information Engineering and Applications (IEA) 2012: Volume 3.Springer, 2013, pp. 375–382.
[9] J. Shalf, “The future of computing beyond moore’s law,” Philosophical Transactions of the Royal Society A, vol. 378, no. 2166, p. 20190061, 2020.
[10] J. d. J. G. Herrera and J. F. B. Vega, “Network functions virtualization: A survey,” IEEE Latin America Transactions, vol. 14, no. 2, pp. 983–997, 2016.
[11] D. Intel, “Data plane development kit,” 2014, https://www.dpdk.org/ [Accessed: July-2023].
[12] B. Goode, “Voice over internet protocol (voip),” Proceedings of the IEEE, vol. 90, no. 9,pp. 1495–1517, 2002.
[13] P. Gepner and M. Kowalik, “Multi-core processors: New way to achieve high system performance,” in International Symposium on Parallel Computing in Electrical Engineering(PARELEC’06), 2006, pp. 9–13.
[14] 程宇閎 et al., “設計與實作-高效能網路功能虛擬化平台,” Ph.D. dissertation, 2022.
[15] S. Bakhtiari, R. Safavi-Naini, J. Pieprzyk et al., “Cryptographic hash functions: A survey,”Citeseer, Tech. Rep., 1995.
[16] A. Rubini and J. Corbet, Linux device drivers. ” O’Reilly Media, Inc.”, 2001.
[17] Huawei, “Load balancing hash algorithms,”https://support.huawei.com/enterprise/en/doc/EDOC1100086965 [Accessed: July-2023].
[18] R. Uhlig, G. Neiger, D. Rodgers, A. L. Santoni, F. C. Martins, A. V. Anderson, S. M.
Bennett, A. Kagi, F. H. Leung, and L. Smith, “Intel virtualization technology,” Computer,vol. 38, no. 5, pp. 48–56, 2005.
[19] “European telecommunications standards institute,” https://www.etsi.org/technologies/nfv [Accessed: July-2023].
[20] “Nfv(22)000098 nfv release 4 description v0 4 0,” https://docbox.etsi.org/ISG/NFV/Open/Other/ReleaseDocumentation [Accessed: July-2023].
[21] W. Zhang, G. Liu, W. Zhang, N. Shah, P. Lopreiato, G. Todeschi, K. Ramakrishnan,
and T. Wood, “Opennetvm: A platform for high performance network service chains,”in Proceedings of the 2016 workshop on Hot topics in Middleboxes and Network Function Virtualization, 2016, pp. 26–31.
[22] J. Xie, M. Miao, F. Ren, W. Cheng, R. Shu, and T. Zhang, “Overload detecting in high performance network i/o frameworks,” in 2016 IEEE 18th International Conference on High Performance Computing and Communications; IEEE 14th International Conferenceon Smart City; IEEE 2nd International Conference on Data Science and Systems (HPC-C/SmartCity/DSS). IEEE, 2016, pp. 999–1006.
[23] M. Vesovi´c, A. Smiljani´c, and M. Tomaˇsevi´c, “Speeding up ip lookup procedure in software routers by means of parallelization,” in 2016 24th Telecommunications Forum (TELFOR).IEEE, 2016, pp. 1–4.
[24] L. Rizzo, “netmap: a novel framework for fast packet i/o,” in 21st USENIX Security Symposium (USENIX Security 12), 2012, pp. 101–112.
[25] S. Han, K. Jang, K. Park, and S. Moon, “Packetshader: a gpu-accelerated software router,”ACM SIGCOMM Computer Communication Review, vol. 40, no. 4, pp. 195–206, 2010.
[26] K. Wiles, “Traffic generator powered by dpdk,” 2019, https://git.dpdk.org/apps/pktgen-dpdk/ [Accessed: July-2023].
[27] INTEL, “What is hyper-threading?” https://www.intel.com/content/www/us/en/gaming/resources/hyperthreading.html [Accessed: July-2023].
[28] DPDK, “rte cycles.h file reference,” https://doc.dpdk.org/api/rte cycles8h.html [Accessed: July-2023].