Shuqiang Zhang

Dynamic Traffic Grooming in Elastic Optical Networks

Spectrum elastic optical networks support flexible central frequency and spectrum assignment for lightpaths. When provisioning a new connection in an elastic optical network that allows traffic grooming, the control plane has to solve two problems: the electrical-layer routing and optical-layer routing and spectrum assignment (RSA). The electrical-layer routing determines how to route the new connection through a combination of new and existing lightpaths, while the optical-layer RSA decides how to establish new lightpaths under the spectrum-continuity constraint. The flexibility (e.g., bandwidth variability of lightpaths) provided by elastic optical networks makes it suitable for accommodating dynamic traffic. It is important and challenging to exploit the full potential of the flexibility when dealing with the above two problems. In this study, we propose a multi-layer auxiliary graph to jointly solve the electrical-layer routing and optical-layer RSA. Various traffic-grooming policies (objectives) can be achieved by properly adjusting the edge weights in the auxiliary graph. Also, we propose a spectrum reservation scheme that can efficiently utilize the bandwidth variability of lightpaths by reserving bandwidth for non-fully utilized lightpaths and grooming future connections onto them. We show that there is a tradeoff among different traffic-grooming policies, and the spectrum reservation scheme can be easily incorporated into various traffic-grooming policies and lead to a significant reduction in operational expenditure (OPEX) and better spectrum efficiency.

Greening the cloud using renewable-energy-aware service migration

Cloud computing is the new paradigm of operation in information technology. While cloud computing infrastructures have benefits, their energy consumption is becoming a growing concern. Data centers, which are used to provide the infrastructure and resource pool for cloud computing, consume a large amount of energy. Future energy consumption predictions of these data centers are even bigger concerns. To reduce this energy consumption, and hence the carbon footprint and greenhouse gas emission of cloud computing, and information technology in general, energy-efficient methods of operation have to be investigated and adopted. In addition, renewable energy usage in place of non-renewable can also reduce carbon emission. However, due to its intermittency and volatility, renewable energy cannot be used to its full potential. In this study, we introduce the renewable-energy- aware cloud service and virtual machine migration to relocate energy demand using dynamic and flexible cloud resource allocation techniques, and help overcome the challenges of renewable energy. Results from a U.S.-wide cloud network infrastructure show that, using simple migration techniques, up to 30 percent nonrenewable energy can be replaced by renewable energy, while consuming only a small amount of extra resources and energy to perform demand relocation.

Software-defined optical networks (SDONs): a survey

This paper gives an overview of software-defined optical networks (SDONs). It explains the general concepts on software-defined networks (SDNs), their relationship with network function virtualization, and also about OpenFlow, which is a pioneer protocol for SDNs. It then explains the benefits and challenges of extending SDNs to multilayer optical networks, including flexible grid and elastic optical networks, and how it compares to generalized multi-protocol label switching for implementing a unified control plane. An overview on the industry and research efforts on SDON standardization and implementation is given next, to bring the reader up to speed with the current state of the art in this field. Finally, the paper outlines the benefits achieved by SDONs for network operators, and also some of the important and relevant research problems that need to be addressed.

Energy-efficient dynamic provisioning for spectrum elastic optical networks

Spectrum elastic optical networks support flexible central frequency and spectrum assignment for lightpaths. In this paper, we investigate energy-efficient dynamic provisioning for such networks. When provisioning a connection, the routing problems in both electrical layer (routing over multiple lightpaths) and optical layer (routing over multiple fibers) have to be addressed. Also, the control plane has to determine the optical-layer spectrum assignment considering the spectrum-continuity constraint as well as the bandwidth variability of transponders. We adopt a novel auxiliary graph based on which a new dynamic provisioning policy called Time-Aware Provisioning with Bandwidth Reservation (TAP-BR) is proposed. TAP-BR incorporates two important factors to facilitate energy-efficient provisioning: time awareness and bandwidth reservation. We compare TAP-BR with previously-proposed dynamic provisioning policies and show that TAP-BR can save significant amount of energy and make efficient use of spectrum resources.

Network virtualization over WDM and flexible-grid optical networks☆

Abstract Network virtualization can eradicate the ossification of the Internet and stimulate innovation of new network architectures and applications. Optical networks are ideal substrates for provisioning high-bandwidth virtual-network services. In this study, we investigate the problem of network virtualization over both WDM and flexible-grid optical networks by formulating the problems as mixed integer linear programs (MILP). Two heuristics, namely MaxMapping and MinMapping , are developed for each kind of network to solve the problem quickly but suboptimally. Numerical examples show that MinMapping consumes fewer spectrum resources than MaxMapping and performs very close to the optimal results derived by the MILP in both kinds of optical networks, by exploring the opportunities of traffic grooming. Also, it is verified that flexible-grid optical networks can be more spectrum efficient than WDM networks as the substrate for network virtualization.

Evolving Traffic Grooming in Multi-Layer Flexible-Grid Optical Networks With Software-Defined Elasticity

Recently, flexible grid and elastic-rate transponders have emerged as evolutionary technologies to satisfy the ever-increasing demand for higher spectrum efficiency and operational flexibility. In this study, we first briefly review the evolution of traffic grooming from SONET/SDH to currently-deployed WDM networks, and summarize the essence of the current traffic-grooming paradigm based on electronic circuit or packet switching and multi-layer collaboration. Then, the role of traffic grooming in flexible-grid and elastic-rate optical networks is re-examined. The impact of some new optical-layer technologies on traffic-grooming paradigm is discussed. Particularly, sliceable optical layer based on sliceable transponders and BV-ROADM is identified as a novel technology that could impact the future grooming paradigm by offloading considerable amount of traffic and part of electronic grooming function to the optical layer. We propose two novel network architectures based on sliceable optical layer and numerically compare them with the traditional packet-over-optical network architecture. It is found that packet-over-sliceable network architecture consumes the fewest transponders and at the same time achieves either the “lowest-possible” latency or least spectrum usage. Finally, traffic grooming, which involves multi-layer resource orchestration, should be controlled by software with a centralized view of the network to accommodate the dynamic requirements of applications.

Network virtualization over WDM networks

Network virtualization can eradicate the ossification of the Internet and stimulate innovation of new architectures and applications. Optical networks are ideal substrates for provisioning high-bandwidth virtual-network services. We investigate the problem of network virtualization over WDM networks by formulating a mixed integer linear program (MILP). Two heuristics, MaxMapping and MinMapping, are developed to solve the problem quickly but suboptimally. Numerical examples show that MinMapping consumes fewer wavelengths than MaxMapping and performs close to the optimal results of the MILP, by exploring the opportunities of traffic grooming.

Heterogeneous bandwidth provisioning for virtual machine migration over SDN-enabled optical networks

Virtual machine migration in cloud-computing environments is an important operational technique, and requires significant network bandwidth. We demonstrate that heterogeneous bandwidth (vs. homogeneous bandwidth) for migration reduces significant resource consumption in SDN-enabled optical networks.

Adopting hybrid CDN-P2P in IP-over-WDM networks: An energy-efficiency perspective

Content-based services are gaining tremendous popularity recently with the proliferation of a wide range of contents. Traditionally, hosting and distribution of these ubiquitous contents are accomplished using content distribution networks (CDNs). CDNs distribute contents in multiple hosting locations for load balancing, availability, and easy access. However, scalability and cost of CDNs are becoming major challenges due to the increase of content-based services. To combat these problems, peer-to-peer (P2P) systems are being considered in combination with CDNtocomplement each other in ahybrid system. Research works and practical deployments have shown that hybrid CDN-P2P systems are feasible, inexpensive, and scalable. However, P2P-based content distribution systems are known to consume more system-wide energy than traditional CDNs. Hence, energy efficiency of such a hybrid system must be taken into consideration before wider implementation. In our study, we analyze the energy consumption of a hybrid CDN-P2P system in an IP-over-WDM network and show that, due to its hybrid nature, it is possible to exploit the P2P system to significantly reduce load on the CDN and at the same time constrain-and in some cases reduce - energy consumption. We propose energy-aware as well as energy-unaware content source selection mechanisms to help service providers achieve this win-win situation. Using simulations, we show that, for example, in case of a moderately popular content the server load can be reduced by 30%-40% while reducing the system-wide energy consumption by 10%-20% whencompared with a traditional CDN.

When and how should the optical network be upgraded to flex grid

This paper addresses the problem of gradual migration from fixed grid to flex grid. We discuss how to perform spectrum assignment in a mixed fixed/flex network, and we propose and compare three upgrade strategies under two traffic models in four different scenarios.