Software-Defined Network Boot Architectures For Hyperscale Data Centers
DOI:
https://doi.org/10.29284/vqewwn08Keywords:
Software-Defined Networking (SDN); Hyperscale Data Centers; Network Boot Architectures; Programmable Data Planes; Secure Attestation.Abstract
The modern cloud computing, artificial intelligence infrastructure, and distributed large-scale services rely on hyperscale data centers. With the number of nodes growing into the hundreds of thousands, the boot procedure has now become a policy-managed workflow whose execution is driven by software-defined network (SDN) fabrics instead of firmware. This paper presents a technical review of software-defined network boot architectures for hyperscale data centres, incorporating the insights of control-plane scalability, programmable data-plane enforcement, secure bootstrapping, and large-scale provisioning systems. A systematic analysis scheme was used to consider peer-reviewed sources related to SDN controller architecture, programmable switching platform, distributed attestation, and zero-touch provisioning framework. Reported metrics synthesis was done in a quantitative way to investigate control-plane reaction latency, rule installation overhead, packet loss during correlated activation, and delay increment due to attestation. Statistical analysis has shown that distributed clusters of controllers can substantially address the issue of scale-induced latency explosion and programmable data-plane isolation can be used to reduce packet loss during the boot phase by nearly a factor of three. Conversely, centralized attestation architectures exhibit nonlinear delay under high concurrency which underscores the need of hierarchical or distributed trust validation models. The results indicate that the cross-layer interactions and not the subsystem efficiencies dictate the performance of the boot. To date, little has been done in terms of empirical testing in real hyperscale configurations and there are no standardized measures of boot performance. Architectural optimization including control-plane elasticity, deterministic forwarding isolation and distributed attestation can be realized significantly to achieve the total provisioning latency however scalability limits remain under correlated boot storms. The validation trace, cross-layer analytical modeling, energy conscious boot scheduling and formal verification of dynamic boot-state transitions should be the priorities in future research. Software-defined boot architectures are positioned as critical enablers of secure, elastic and composable hyperscale infrastructure, and further interdisciplinary research is recommended.
References
[1] Armbrust, M., Fox, A., Griffith, R., Joseph, A. D., Katz, R., Konwinski, A., & Zaharia, M. (2010). A view of cloud computing. Communications of the ACM, 53(4), 50–58.
[2] Buyya, R., Broberg, J., & Goscinski, A. (2011). Cloud computing: Principles and paradigms. Wiley.
[3] Barroso, L. A., Clidaras, J., & Hölzle, U. (2013). The datacenter as a computer (2nd ed.). Morgan & Claypool.
[4] Kreutz, D., Ramos, F. M., Verissimo, P. E., Rothenberg, C. E., Azodolmolky, S., & Uhlig, S. (2015). Software-defined networking: A comprehensive survey. Proceedings of the IEEE, 103(1), 14–76.
[5] Nunes, B. A., Mendonca, M., Nguyen, X. N., Obraczka, K., & Turletti, T. (2014). A survey of SDN. IEEE Communications Surveys & Tutorials, 16(3), 1617–1634.
[6] Scott-Hayward, S., O’Callaghan, G., & Sezer, S. (2016). SDN security survey. IEEE Communications Surveys & Tutorials, 18(1), 623–654.
[7] McKeown, N., et al. (2008). OpenFlow. ACM SIGCOMM CCR, 38(2), 69–74.
[8] Gude, N., et al. (2008). NOX. ACM SIGCOMM CCR, 38(3), 105–110.
[9] Tootoonchian, A., & Ganjali, Y. (2010). HyperFlow. INM/WREN.
[10] Koponen, T., et al. (2010). Onix. USENIX OSDI.
[11] Berde, P., et al. (2014). ONOS. HotSDN.
[12] Medved, J., et al. (2014). OpenDaylight. IEEE ICNP Workshop.
[13] Dixit, A., Hao, F., Mukherjee, S., Lakshman, T. V., & Kompella, R. (2013). Elastic distributed SDN controllers. ACM SIGCOMM CCR, 43(4), 7–12.
[14] Phemius, K., & Bouet, M. (2013). Scalable SDN controllers. IEEE CNSM.
[15] Al-Fares, M., Loukissas, A., & Vahdat, A. (2008). A scalable data center network architecture. ACM SIGCOMM CCR, 38(4), 63–74.
[16] Jain, S., et al. (2013). B4 WAN. ACM SIGCOMM CCR, 43(4), 3–14.
[17] Roy, A., et al. (2015). Datacenter network insights. ACM SIGCOMM.
[18] Benson, T., Akella, A., & Maltz, D. A. (2010). Data center traffic characteristics. ACM IMC.
[19] Bosshart, P., et al. (2014). P4 language. ACM SIGCOMM CCR, 44(3), 87–95.
[20] Sivaraman, A., et al. (2016). Packet transactions. ACM SIGCOMM.
[21] Dang, H. T., et al. (2015). NetFPGA SUME. IEEE Micro, 35(5), 32–41.
[22] Alizadeh, M., et al. (2014). CONGA. ACM SIGCOMM.
[23] Yu, M., et al. (2010). DIFANE. ACM SIGCOMM.
[24] Gao, P. X., Narayan, A., Menache, I., & Ramakrishnan, K. K. (2016). Resource disaggregation. USENIX NSDI.
[25] Sefraoui, O., et al. (2012). OpenStack. International Journal of Computer Applications, 55(3), 38–42.
[26] Verma, A., et al. (2015). Borg system. EuroSys.
[27] Mao, M., & Humphrey, M. (2012). VM startup time. IEEE CLOUD.
[28] Francillon, A., Nguyen, Q., Rasmussen, K. B., & Tsudik, G. (2014). Remote attestation. Designs, Codes and Cryptography, 72(1), 51–70.
[29] Sadeghi, A.-R., & Stüble, C. (2004). Property-based attestation. NSPW.
[30] Trusted Computing Group. (2019). TPM 2.0 library specification.
[31] Ibrahim, A., et al. (2017). Scalable attestation. IEEE TDSC.
[32] Costan, V., & Devadas, S. (2016). Intel SGX explained. IACR ePrint.
[33] Kazemian, P., Varghese, G., & McKeown, N. (2012). Header space analysis. USENIX NSDI.
[34] Lantz, B., Heller, B., & McKeown, N. (2010). Mininet. HotNets.
[35] Riley, G. F., & Henderson, T. R. (2010). ns-3 simulator. Springer.
[36] Mao, H., Alizadeh, M., Menache, I., & Kandula, S. (2016). Deep RL resource management. ACM HotNets.
[37] Dorri, A., et al. (2017). Blockchain for IoT security. IEEE IoT Journal, 4(5), 1815–1826.
[38] Greenberg, A., et al. (2009). VL2. ACM SIGCOMM.
[39] Kandula, S., et al. (2009). Nature of data center traffic. ACM IMC.
[40] Hong, C.-Y., et al. (2013). Achieving high utilization. ACM SIGCOMM.
[41] Farrington, N., et al. (2010). Helios. ACM SIGCOMM.
[42] Vahdat, A., et al. (2010). Scale-out networking. IEEE Micro.
[43] Handigol, N., et al. (2012). Reproducible network experiments. CoNEXT.
[44] Feamster, N., Rexford, J., & Zegura, E. (2014). SDN research roadmap. ACM CCR.
[45] Curtis, A. R., et al. (2011). DevoFlow. ACM SIGCOMM.
[46] Voellmy, A., & Wang, J. (2012). Scalable OpenFlow. ACM SIGCOMM.
[47] Wang, G., et al. (2014). Load balancing SDN. IEEE INFOCOM.
[48] Rexford, J. (2013). SDN principles. ACM Queue.
[49] Shin, S., et al. (2013). AVANT-GUARD. ACM CCS.
[50] Porras, P., et al. (2015). SDN security enforcement. NDSS.
[51] Sezer, S., et al. (2013). SDN challenges. IEEE Communications Magazine.
[52] Intel Corporation. (1999). PXE specification.
[53] Baset, S. A., et al. (2012). Resource management as a service. IEEE Network.
[54] Fonseca, P., et al. (2012). Resilient OpenFlow networking. IEEE NOMS.
[55] Jain, R. (1991). The art of computer systems performance analysis. Wiley.
[56] Paxson, V. (1997). End-to-end Internet packet dynamics. IEEE/ACM TON.
[57] Floyd, S., & Jacobson, V. (1993). Random early detection gateways. IEEE/ACM TON.
[58] Fall, K., & Varadhan, K. (2011). ns Manual. UC Berkeley.
[59] Peterson, L., & Davie, B. (2011). Computer networks: A systems approach. Morgan Kaufmann.
[60] Tanenbaum, A. S., & Wetherall, D. (2011). Computer networks (5th ed.). Pearson.
[61] Stallings, W. (2013). Data and computer communications. Pearson.
[62] Cisco Systems. (2016). Cisco SDN architecture white paper.
[63] ETSI. (2014). Network functions virtualization (NFV) white paper.
[64] Open Networking Foundation. (2012). SDN architecture overview.
[65] Amazon Web Services. (2018). AWS infrastructure overview.
[66] Google Cloud. (2019). Google data center design white paper.
[67] Microsoft Azure. (2018). Azure architecture guide.
[68] VMware. (2017). NSX network virtualization platform.
[69] Red Hat. (2019). OpenStack architecture guide.
[70] Linux Foundation. (2020). ONAP architecture overview.
[71] Open Compute Project. (2016). Data center hardware design guide.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
