Inter-datacenter (inter-DC) networks are essential for large enterprises to deliver high-quality services to end-users. Since DCs are vulnerable to natural disasters, an inter-DC network operator needs an effective emergency backup plan to evacuate the endangered data out in case of a progressive disaster whose status can be predicted by an early warning system. In this paper, we try to solve the problem of emergency backup in inter-DC networks with progressive disasters. We first utilize the time-expanded network (TEN) approach to model the time-variant inter-DC network during a progressive disaster as a variant TEN (VTEN) and convert the dynamic flow scheduling for emergency backup to a static one. Then, with the VTEN, we formulate an optimization model to maximize the profit from the emergency backup in consideration of data values and resource costs. Although this large-scale optimization can be solved in a distributed way by leveraging the alternation direction method of multipliers (ADMM), we find that one of its subproblems is nontrivial in the distributed setting. We propose a novel inexact ADMM approach to resolve the issue induced by the subproblem, and prove that the proposed algorithm can converge to the optimal solution. The results from extensive simulations confirm that our algorithm is robust and time-efficient, and outperforms several benchmarks in terms of backup profit and running time.

This paper considers the emergency backup in an inter-datacenter (inter-DC) network whose topology is time-variant due to the progress of a disaster. We first transform the dynamic backup into a static flow problem through building a variable time-expanded network (V-TEN). Then, by considering both data utility and resource cost, we formulate an optimization to maximize the backup profit and leverage the alternating direction method of multipliers (ADMM) to design a time-efficient and distributed algorithm. Simulation results show that our ADMM-based algorithm outperforms several existing ones.

Network function virtualization (NFV) is a promising technology to decouple the network functions from dedicated hardware elements, leading to the significant cost reduction in network service provisioning. As more and more users are trying to access their services wherever and whenever, we expect the NFV-related service function chains (SFCs) to be dynamic and adaptive, i.e., they can be readjusted to adapt to the service requests' dynamics for better user experience. In this paper, we study how to optimize SFC deployment and readjustment in the dynamic situation. Specifically, we try to jointly optimize the deployment of new users' SFCs and the readjustment of in-service users' SFCs while considering the trade-off between resource consumption and operational overhead. We first formulate an integer linear programming (ILP) model to solve the problem exactly. Then, to reduce the time complexity, we design a column generation (CG) model for the optimization. Simulation results show that the proposed CG-based algorithm can approximate the performance of the ILP and outperform an existing benchmark in terms of the profit from service provisioning.

This paper studies the problem of service chain (SC) deployment. Specifically, we try to place virtual network functions (vNFs) on network nodes and connect the vNFs in sequence through link mapping. We start with the offline problem. An integer linear programming (ILP) model is formulated to minimize the total SC deployment cost. With the ILP, we prove that the offline problem is NP-hard and propose a time-efficient heuristic based on affiliation-aware vNF placement. Then, we move to the online problem, and design a forecast-assisted online SC deployment algorithm that includes the prediction of future vNF requirements. Simulation results show that the online algorithms can reduce the blocking probability of SC requests and increase the service provider's profit from SC deployment effectively.

For an inter-datacenter (inter-DC) network that carries multiple cloud applications, tasks may arise in the DCs and need to transfer data to others with different latency requirements. Therefore, it is desired that a highly efficient online scheduling algorithm could be developed to consider the request admission, routing selection and bandwidth allocation for the data-transfers jointly. In this work, we investigate this problem, and propose an online scheduling algorithm, namely, GlobeAny, to maximize the time-average profit from provisioning the requests. The proposed algorithm leverages Lyapunov optimization techniques and can achieve arbitrarily approaching to the optimal value within O(1/V) gap. We also show that by adjusting the application weights, it can provide differentiated services to requests with different latency requirements.

It is known that with a hybrid cloud, a multimedia cloud service provider (MCSP) can quickly extend its services to multiple geographical locations with quality-of-service (QoS) guarantees. Meanwhile, to maximize its profit, the MCSP needs an online management mechanism to operate the hybrid cloud efficiently. In this paper, we study how to maximize an MCSP's profit from provisioning multimedia services to geographically distributed users with a hybrid cloud. We first design a service provisioning model to manage the resources in the hybrid cloud. Here, in order to make the model practical and address the different situations in private and public clouds, we consider different time granularities for resource reservations. Then, we leverage the Lyapunov optimization technique to maximize the profit of MCSP and propose an online algorithm that can manage the hybrid cloud in the distributed manner. Specifically , the algorithm determines the access control and routing of each multimedia service request, and allocates the resources in the hybrid cloud accordingly. We also apply the ε-persistent technique to ensure that the worst-case latency of the provisioned requests is bounded. Finally, the proposed algorithm is evaluated with extensive simulations using both synthetical and real traces. Simulation results indicate that the algorithm can manage the hybrid cloud efficiently and maximize the profit of MCSP.

We optimize the disaster backup in multi-datacenter (multi-DC) cloud systems and design disaster-aware algorithms to realize rapid backup with the objective of minimizing the backup window for all the DCs in the network. A mixed integer linear programming (MILP) model is first formulated to optimize the backup processes of all production DCs jointly. We then develop three heuristics that use the one-step or two-step approaches for the selection of backup DCs and the calculation of backup routing paths. Simulation results show that the Two-Step algorithm can achieve the shortest backup window with the lowest operation complexity.

We develop a novel traffic scheduling algorithm based on a master-slave linecard (LC) configuration to improve the energy-efficiency of hybrid fiber-coaxial (HFC) networks. The algorithm forwards packets to the master or slave LC adaptively according to the traffic load, and toggles the LCs between working and sleeping modes for energy-saving. To optimize the algorithm's key parameters, we model the system using a two-dimensional Markov process and derive the analytical expressions of several performance metrics, including average packet delay, LC switching frequency, and energy efficiency improvement. We then verify the theoretical analysis with numerical simulations using the Monte Carlo method. Both the theoretical and simulation results indicate that the proposed algorithm can achieve significant energy efficiency improvement, while keeping the average packet delay and LC switching frequency low.

Hybrid Fiber Coaxial (HFC) networks support broadband Internet access with the existing cable TV systems. Recent advances on HFC networks have demonstrated effective improvements on the customers' access speeds with the channel-bonding technology. In this paper, we develop a novel energy-efficient traffic scheduling algorithm for the HFC networks that support channel bonding. The proposed algorithm is compliant with the newly-released DOCSIS 3.0 standard. We come up with a system model of the channel-bonding transmitters (TXs) on a Cable Modem (CM), and then define several operation modes for them. At the beginning of each scheduling cycle, the proposed algorithm adjusts the TXs' operation modes based on the traffic status. Both analytical analysis and numerical simulations are then developed to investigate the energy-saving and delay introduced by the algorithm. The results on energy-saving indicate that the energy-consumption of the TXs scales almost linearly with the input traffic load and effective energy-saving can be achieved. We also investigate the tradeoff between the energy-saving and the average delay to optimize the parameters for the energy-efficient scheduling, and to make sure that the traffic will not experience significant delay increase due to the energy-saving operations.