Hongfang Yu

Path-based routing provisioning with mixed shared protection in WDM mesh networks

In this paper, the authors focus on studying the problem of survivable routing provisioning to prevent single link failure in wavelength-division-multiplexing (WDM) mesh networks, and propose a novel protection scheme called mixed shared path protection (MSPP). With MSPP, the authors define three types of resources: 1) primary resources that can be used by primary paths; 2) spare resources that can be shared by backup paths; and 3) mixed resources that can be shared by both the primary and the backup paths. In the proposed protection scheme, each connection is assigned a primary path and a link disjoint backup path. Differing from pervious protection schemes, MSPP allows some primary paths and backup paths to share the common mixed resources if the corresponding constraints can be satisfied. In this paper, the authors consider three types of path-based protection schemes, i.e., dedicated path protection (DPP), shared path protection (SPP), and MSPP, and evaluate their performance for both the static and the dynamic provisioning problems. Simulation results show that MSPP outperforms DPP and SPP

Cost Efficient Design of Survivable Virtual Infrastructure to Recover from Facility Node Failures

As network virtualization becomes popular, the problem of efficiently mapping a virtual infrastructure (VI) over a substrate network while guaranteeing its survivability in the event of failures becomes increasingly important. In this paper, we study the survivable VI mapping problem to recover from facility node failures. We develop two solutions namely the 1-redundant scheme and the K-redundant scheme for surviving facility node failures while minimizing network resource costs. We also model the two schemes as a MILP problem and propose efficient heuristics based on the MILP formulations. We compare the efficiency of our solutions using simulation under various performance metrics.

Survivable Virtual Infrastructure Mapping in a Federated Computing and Networking System under Single Regional Failures

As virtualization becomes more and more popular, how to guarantee survivability of a virtual infrastructure (VI) over a wide-area optical network is increasingly important. In this paper, we approach the problem of survivable VI mapping (SVIM) from a few unique perspectives. One of the most distinguishing perspectives is that a large-scale regional failure could destroy one or more facility nodes to which some VI nodes are mapped. Accordingly, redundant facility nodes at different geographical locations and redundant optical connections have to be provisioned such that the VI can still be mapped after the failure. Another distinguishing perspective is that with failure-dependent protection, the SVIM problem can be decomposed into several instances of the basic non-survivable VI mapping (NSVIM) problem, whose solution permits effective sharing of the redundant resources among all failures. In this paper, we first formulate the minimum-cost SVIM problem using mixed integer linear programming (MILP). We then propose an efficient heuristic solution to NSVIM, based on which two novel heuristic SVIM algorithms called Separate Optimization with Unconstrained Mapping (SOUM) and Incremental Optimization with Constrained Mapping (IOCM). Simulations are performed to study and compare the performance of the MILP and heuristics.

A cost efficient framework and algorithm for embedding dynamic virtual network requests

Cloud computing is a novel paradigm that enables transparent resource sharing over the Internet. With cloud computing users access applications/services and infrastructure resources using thin clients without knowing the actual location or characteristics of the resources. These applications are typically hosted and run on servers in interconnected data centers. The task or application request from the same or different users can be abstracted as virtual network (VN) requests, which are supported by the same underlying substrate network and thus share its resources. Thus, efficient mapping techniques that intelligently use the substrate network resources are important. Current research only considers the case when the VN requests are static. However, user demands and the corresponding VN requests can change dynamically. In this paper, we address the issue of how to optimally reconfigure and map an existing VN while the VN request changes. We first model this problem as a mathematical optimization problem with the objective of minimizing the reconfiguration cost by using mixed integer linear programming. Since the optimal problem is NP-hard we also propose heuristic algorithms for solving it efficiently. We validate and evaluate our framework and algorithms by conducting extensive simulations on different realistic networks under various scenarios, and by comparing with existing approaches. Our simulation results show that our approach outperforms existing solutions.

Dynamic Grooming Algorithms for Survivable WDM Mesh Networks

Within a WDM grooming mesh network and under the constraints of the number of transceivers per node and wavelength continuity, we propose a novel dynamic grooming graph which models the number of transceivers per node in addition to the usage of wavelength and bandwidth resources. Based on the grooming graph, we first propose a dynamic traffic-grooming algorithm called integrated grooming algorithm (IGA). And we also propose two dynamic survivable traffic-grooming algorithms, which are called protection per lightpath traffic-grooming algorithm (PPL) and protection per connection traffic-grooming algorithm (PPC). These algorithms are evaluated via simulations.

Survivable Virtual Network Design and Embedding to Survive a Facility Node Failure

As virtualization is becoming a promising way to support various emerging application, provisioning survivability to requested virtual networks (VN) in a resource efficient way is important. In this paper, we investigate the survivable VN embedding (SVNE) problem from a new perspective. First, we consider the failure dependent protection (FDP) in which each primary facility node would have a different backup facility node, as opposed to the Failure Independent Protection (FIP) which has been studied before, in order to provide the same degree of protection against a single node failure with less substrate resources. Secondly, we enhance the VN with additional computing and communication resources and design the Enhanced VN (or EVN) before embedding it to the substrate in order to further reduce the amount of substrate resources needed to survive a single facility node failure. The work is the first that combines the FDP with EVN design (FD-EVN) to explore a resource efficient solution to the SVNE problem. After presenting a binary quadratic programming (BQP) formulation of the FD-EVN design problem and a Mixed Integer Linear Programming (MILP) formulation of the EVN embedding (EVNE) problem, we propose two heuristic algorithms for FD-EVN design, as well as an EVNE algorithm that explores primary and backup substrate resources sharing. Simulations are conducted to evaluate the performance of the solutions to the BQP/MILP formulation when possible, and the heuristics. The proposed FD-EVN approach has shown to be resource efficient and in particular, outperform other approaches in terms of request acceptance ratio and embedding cost, although as a tradeoff, requiring more service migration after failures.

On progressive network recovery after a major disruption

A major disruption may affect many network components and significantly lower the capacity of a network measured in terms of the maximum total flow among a set of source-destination pairs. Since only a subset of the failed components may be repaired at a time due to e.g., limited availability of repair resources, the network capacity can only be progressively increased over time by following a recovery process that involves multiple recovery stages. Different recovery processes will restore the failed components in different orders, and accordingly, result in different amount of network capacity increase after each stage. This paper aims to investigate how to optimally recover the network capacity progressively, or in other words, to determine the optimal recovery process, subject to limited available repair resources. We formulate the optimization problem, analyze its computational complexity, devise solution schemes, and conduct numerical experiments to evaluate the algorithms. The concept of progressive network recovery proposed in this paper represents a paradigm-shift in the field of resilient and survivable networking to handle large-scale failures, and will motivate a rich body of research in network design and other applications.

On Finding Feasible Solutions With Shared Backup Resources for Surviving Double-Link Failures in Path-Protected WDM Mesh Networks

In this paper, we investigate the problem of dynamic survivable routing in wavelength-division-multiplexing optical networks and propose a new path-protection algorithm called Enhanced Shared Backup Paths Protection (ESBPP) to provide the complete survivability for double-link failures. In order to improve the resource-utilization ratio, we present the new rule of sharing backup resources that can yield better resource-utilization ratio than previous methods. In order to overcome the problem of trap paths, we develop a new routing method to find the feasible solution with three link-disjoint paths for each connection request. Compared to previous algorithms for double-link failures, ESBPP can obtain higher resource-utilization ratio and lower blocking probability

A new technique for efficient live migration of multiple virtual machines

Datacenter virtualization technologies have attracted a lot of attention to enable various cloud computing services and to facilitate virtual machine (VM) migration. VM migration can help service providers to achieve the goals of saving energy, enhancing resource efficiency and quality of service (QoS). In order to ensure the QoS, the migration time and the downtime of VM should be considered while implementing the VM migration. Most researches focus on the issue of single VM migration by using the post-copy migration strategy or pre-copy migration strategy. However, there is few research that focuses on the problem of live migration for multiple VMs. Therefore, in this paper, we first propose an improved serial migration strategy and introduce the post-copy migration scheme into it. We then propose the m mixed migration strategy that is based on the improved serial migration strategy and the parallel migration strategy. Furthermore, we develop queuing models (i.e., the M/M/C/C and the M/M/C queuing models) to quantify performance metrics, such as the blocking ratio and average waiting time of each migration request. We evaluate the performance of the proposed migration strategy by conducting mathematical analysis, the numerical results show that our proposed strategy outperforms the existing approach. We propose an improved serial migration strategy for migrating multiple VMs.We present the m mixed migration strategy for improving the time efficiency.For a given tolerable downtime, we calculate m to minimize the migration time.We evaluate the effectiveness of the proposed approach by using queuing models.

A Cost Efficient Design of Virtual Infrastructures with Joint Node and Link Mapping

The virtualization of both servers and substrate networks is a promising technique that will enable the future Internet architecture to support a variety of cloud computing services and architectures. The problem of efficiently mapping a virtual infrastructure (VI) over a wide-area optical network is a key problem in provisioning Infrastructure (and Network) as a Service over multiple datacenters. In this paper, we study the problem of cost efficient VI mapping and propose a novel approach called the virtual infrastructure mapping algorithm (VIMA) that jointly considers node mapping and link mapping. Such a close coordination between the link and node mapping stages increases the efficiency of the VIMA algorithm. We compare the performance of our VIMA algorithm with other VI mapping algorithms such as NSVIM, vnmFlib and R-ViNE under various performance metrics using simulation. Our simulation results and analysis show that the VIMA algorithm outperforms the other algorithms in terms of VI mapping costs and path length of VI links.