Distributed in-network cloud architecture is a promising solution to efficiently host next generation internet-of-things (IoT) services. With the rapid increase of IoT devices and applications, the backhaul or backbone networks, which transmit IoT traffic to various in-network clouds, will experience a predicted explosion in the volume of carried traffic. To guarantee the QoS of IoT cloud services and improve the network performance, it is crucial for network operator to implement efficient routing optimization strategies for IoT traffic. As a promising networking paradigm, software-defined networking (SDN) has flexible and programmable control capability for fine-grained flows. The emergence of SDN paves a way for implementing high-performance routing optimization in networks. In SDN networks, the routing strategies are realized through flow rules, which are usually stored in TCAM with very limited capacity. However, the number of IoT flows are enormous. Thus, in this paper, we address the routing optimization problem in SDN-based IoT with TCAM capacity constraint. We first formulate the problem as a mixed integer linear programming problem and prove the problem is NP-hard. Then to solve the problem efficiently, we propose several approximate algorithms, which solve the problem in two stages. In the first stage, the algorithms calculate the routing strategies for flows without considering the TCAM capacity constraint. To meet the TCAM capacity constraint, the algorithms using different strategies to adjust the paths of some flows in the second stage. Extensive simulations are conducted on both real ISP and synthetic topologies to evaluate the performance of the algorithms. The simulation results verify that the algorithms can achieve promising load balancing performance in SDN-based IoT, where the capacity of TCAM in SDN switches is very limited.

中文翻译：

---

具有TCAM容量约束的基于SDN的物联网云服务路由优化

分布式网络内云架构是有效托管下一代物联网 (IoT) 服务的有前途的解决方案。随着物联网设备和应用的快速增加，将物联网流量传输到各种网络内云的回程或骨干网络将经历预计的承载流量爆炸式增长。为了保证物联网云服务的 QoS 并提高网络性能，网络运营商对物联网流量实施高效的路由优化策略至关重要。作为一种很有前途的网络范式，软件定义网络（SDN）对细粒度流具有灵活和可编程的控制能力。SDN的出现为在网络中实现高性能路由优化铺平了道路。在SDN网络中，路由策略是通过流规则实现的，流规则通常存储在容量非常有限的TCAM中。然而，物联网流量的数量是巨大的。因此，在本文中，我们解决了具有 TCAM 容量约束的基于 SDN 的 IoT 中的路由优化问题。我们首先将问题表述为混合整数线性规划问题，并证明该问题是 NP-hard 问题。然后为了有效地解决问题，我们提出了几种近似算法，它们分两个阶段解决问题。在第一阶段，算法在不考虑 TCAM 容量约束的情况下计算流的路由策略。为了满足TCAM容量约束，算法在第二阶段使用不同的策略来调整一些流的路径。在真实的 ISP 和合成拓扑上进行了广泛的模拟，以评估算法的性能。仿真结果验证了该算法可以在基于 SDN 的物联网中实现良好的负载平衡性能，其中 SDN 交换机中的 TCAM 容量非常有限。