臺灣博碩士論文加值系統

在資料中心交換機內的微秒脈衝(Microburst)發生於極短的時間內，且在大觀察尺度下難以察覺其存在。先前被提出的相關研究中，有一部分致力於尋找微秒脈衝的偵測方法。然而目前為止卻尚未有紓解這一現象的方式被提出。因此，在這個研究中提出一個實時的微秒脈衝紓解機制 MODFIM。MODFIM將流分段的間隔來達成有序的封包繞道，藉此在網路內部減少封包遺失。MODFIM由三個部分組成：(1) Detour Launcher (2) Packet Sequencer以及(3)Post-detour Handler。利用可程式化交換機的功能，本研究提出的機制可以被完全實作在網路內(而不需要對端主機的修改)。盡作者所知，本研究是第一個可以在實時紓解微秒脈衝，同時能在沒有封包被丟棄的前提下最小化封包亂序情況的研究。

Microbursts in the datacenter network (DCN) last for a very short period in switches and are difficult to be discovered from a course-grained perspective; Some previous works are dedicated to microburst detection in the network and have not yet tried to mitigate microbursts. Therefore, in this paper, a real-time microburst mitigator, MODFIM, is proposed and utilizes orderly detour with flowlet intervals to reduce packet loss in an in-network manner. MODFIM consists of three major components: (1) Detour Launcher, (2) Packet Sequencer and (3) Post-detour Handler. With the aid of programmable switches, our approach can entirely work in the network. To the best of our knowledge, MODFIM is also the first solution capable of mitigating microbursts in real time, meanwhile achieving zero loss and minimizing packet reordering.

摘要 i
ABSTRACT ii
誌謝 iii
List of Figures vi
List of Tables vii
Chapter 1: Introduction 1
Chapter 2: Architecture of MODFIM 7
2.1 Detour Launcher 7
2.2 Packet Sequencer 9
2.3 Post-detour Handler 11
2.4 MODFIM Under Complicated Circumstances 13
2.4.1 Queue-Depth Monitor (QDM) 14
2.4.2 Real-Time Sorter (RTS) 14
Chapter 3: Hardware Limitation on MODFIM 15
3.1 Modification on Detour Launcher 15
3.2 Modification on Packet Sequencer 16
3.3 Modification on Post-detour Handler 17
3.4 Modification on QDM and RTS 17
3.5 Extended Discussion 18
Chapter 4: Hardware Proof of Concept (POC) on P4 Switch 20
4.1 Testbed 20
4.2 Experimental Results 22
4.3 What Could Make MODFIM Fail 27
4.3.1 Packet loss occurs on the link? 27
4.3.2 All neighboring switches are congested? 27
4.3.3 Are 32 bits long enough for encoding sequence numbers? 28
Chapter 5: Related Works 29
5.1 In-network microburst detection 29
5.2 Packet detouring 29
5.3 TCP Flowlet application 30
Chapter 6: Conclusion 31
Bibliography 32

[1] S. Kunniyur and R. Srikant, "End-to-end congestion control schemes: utility functions, random losses and ecn marks", IEEE/ACM Transactions on Networking, vol. 11, no. 5, pp. 689-702, 2003.
[2] M. Alizadeh et al., "Data center TCP (DCTCP)", ACM SIGCOMM Computer Communication Review, vol. 40, no. 4, p. 63, 2010.
[3] H. Wu, J. Ju, G. Lu, C. Guo, Y. Xiong and Y. Zhang, "Tuning ECN for data center networks", Proceedings of the 8th international conference on Emerging networking experiments and technologies - CoNEXT '12, 2012.
[4] D. Shan and F. Ren, "Improving ECN marking scheme with micro-burst traffic in data center networks", IEEE INFOCOM 2017 - IEEE Conference on Computer Communications, 2017.
[5] R. Mittal et al., "TIMELY", Proceedings of the 2015 ACM Conference on Special Interest Group on Data Communication - SIGCOMM '15, 2015.
[6] H. Jiang and C. Dovrolis, "Why is the internet traffic bursty in short time scales?", Proceedings of the 2005 ACM SIGMETRICS international conference on Measurement and modeling of computer systems - SIGMETRICS '05, 2005.
[7] M. Allman and E. Blanton, "Notes on burst mitigation for transport protocols", ACM SIGCOMM Computer Communication Review, vol. 35, no. 2, p. 53, 2005.
[8] T. Benson, A. Anand, A. Akella and M. Zhang, "Understanding data center traffic characteristics", ACM SIGCOMM Computer Communication Review, vol. 40, no. 1, p. 92, 2010.
[9] T. Benson, A. Akella and D. Maltz, "Network traffic characteristics of data centers in the wild", Proceedings of the 10th annual conference on Internet measurement - IMC '10, 2010.
[10] Q. Zhang, V. Liu, H. Zeng and A. Krishnamurthy, "High-resolution measurement of data center microbursts", Proceedings of the 2017 Internet Measurement Conference on - IMC '17, 2017.
[11] D. Shan, F. Ren, P. Cheng, R. Shu and C. Guo, "Micro-Burst in Data Centers: Observations, Analysis, and Mitigations", 2018 IEEE 26th International Conference on Network Protocols (ICNP), 2018.
[12] X. Chen, S. Feibish, Y. Koral, J. Rexford and O. Rottenstreich, "Catching the Microburst Culprits with Snappy", Proceedings of the Afternoon Workshop on Self-Driving Networks - SelfDN 2018, 2018.
[13] R. Joshi, T. Qu, M. Chan, B. Leong and B. Loo, "BurstRadar", Proceedings of the 9th Asia-Pacific Workshop on Systems - APSys '18, 2018.
[14] The P4 Language Consortium. P416 Language Specification. (Nov 2018). [Online]. Available: {https://p4.org/p4-spec/docs/P4-16-v1.1.0-spec.pdf}
[15] The P4 Language Consortium. P416 Portable Switch Architecture. (Nov 2018). [Online]. Available: {https://p4.org/p4-spec/docs/PSA-v1.1.0.pdf}
[16] The P4 Language Consortium. P4 Data-plane Telemetry, In-band Network Telemetry. [Online].Available: {https://github.com/p4lang/p4-applications/blob/master/docs/INT.pdf}
[17] R. Miao, H. Zeng, C. Kim, J. Lee and M. Yu, "SilkRoad", Proceedings of the Conference of the ACM Special Interest Group on Data Communication - SIGCOMM '17, 2017.
[18] N. Katta, M. Hira, C. Kim, A. Sivaraman, and J. Rexford, “Hula,” Proceedings of the Symposium on SDN Research - SOSR 16, 2016.
[19] M. Alizadeh, N. Yadav, G. Varghese, T. Edsall, S. Dharmapurikar, R. Vaidyanathan, K. Chu, A. Fingerhut, V. T. Lam, F. Matus, and R. Pan, “Conga,” Proceedings of the 2014 ACM conference on SIGCOMM - SIGCOMM 14, 2014.
[20] K. Zarifis, R. Miao, M. Calder, E. Katz-Bassett, M. Yu, and J. Padhye, “Dibs,” Proceedings of the Ninth European Conference on Computer Systems - EuroSys 14, 2014.
[21] X. Shi, L. Wang, F. Zhang, K. Zheng, and Z. Liu, “PABO: Congestion mitigation via packet bounce,” 2017 IEEE International Conference on Communications (ICC), 2017.
[22] S. Sinha, S. Kandula, and D. Katabi. ‘’Harnessing TCPs Burstiness using Flowlet Switching,’’ 3rd ACM SIGCOMM Workshop on Hot Topics in Networks (HotNets), 2004.
[23] K. He, E. Rozner, K. Agarwal, W. Felter, J. Carter, and A. Akella, “Presto,” Proceedings of the 2015 ACM Conference on Special Interest Group on Data Communication - SIGCOMM 15, 2015.
[24] M. Al-Fares, S. Radhakrishnan, B. Raghavan, N. Huang, and A. Vahdat. “Hedera: dynamic flow scheduling for data center networks,” Proceedings of the 7th USENIX conference on Networked systems design and implementation – NSDI 10, 2010.