Fangfang Yan

Nonblocking Four-Stage Multicast Network for Multicast-Capable Optical Cross Connects

In this paper, we investigate the four-stage multicast network with only two multicast stages (the second and output stages) being multicast-capable, which is called 4(2 m)-stage multicast network. We first derive the sufficient wide-sense nonblocking (WSNB) and rearrangeable nonblocking (RNB) conditions for the 4(2 m)-stage multicast network. The WSNB and RNB conditions are also presented for multicast request with limited multicast degree. The 4(2 m)-stage multicast network needs at least O(N 3 / 2) crosspoints to be WSNB and RNB, which is a lower bound than the three-stage Clos network with two multicast stages and the same bound with the four-stage network with three multicast stages. Then we design full and partial multicast-capable optical cross connects (MC-OXCs) applying the 4(2 m)-stage multicast network. The proposed WSNB full MC-OXC with two multicast stages is power efficient in reducing splitting loss and has lower multicast cost, compared with the one having three multicast stages. Partial MC-OXC with only one multicast stage is proposed to further reduce the multicast cost and splitting loss. The RNB MC-OXCs with one or two multicast stages provide better performance at the respects of crosspoints, multicast cost, and splitting loss than their WSNB counterparts.

Congestion-Aware Embedding of Heterogeneous Bandwidth Virtual Data Centers With Hose Model Abstraction

Predictable network performance is critical for cloud applications and can be achieved by providing tenants a dedicated virtual data center (VDC) with bandwidth guarantee. Recently, the extended Hose model was applied to the VDC abstraction to characterize the tradeoff between cost and network performance. The acceptability determination problem of a VDC with heterogeneous bandwidth demand was proved to be NP-complete, even in the simple tree topology. In this paper, we investigate the embedding problem for heterogeneous bandwidth VDC in substrate networks of general topology. The embedding problem involves two coupled sub-problems: virtual machine (VM) placement and multipath route assignment. First, we formulate the route assignment problem with linear programming to minimize the maximum link utilization, and provide

Efficient Sharing of Fixed Wavelength Converters in Clos-Type Wavelength Interchanging Cross Connects

K

Wavelength Converted Broadcast-Selective Buffering Contention Resolution in Synchronous WDM OPS Networks

-widest path load-balanced routing with controllable splitting paths. Next, we propose a polynomial-time heuristic algorithm, referred to as the perturbation algorithm, for the VM placement. The perturbation algorithm is congestion-aware as it detects the bandwidth bottlenecks in the placement process and then selectively relocates some assigned VMs to eliminate congestion. Simulation results show that our algorithm performs better in comparison with the existing well-known algorithms: first-fit, next-fit, and greedy, and very close to the exponential-time complexity backtracking algorithm in typical data center network architectures. For the tree substrate network, the perturbation algorithm performs better than the allocation-range algorithm. For the homogeneous bandwidth VDC requests, the perturbation algorithm produces a higher success rate than the recently proposed HVC-ACE algorithm. Therefore, it provides a compromised solution between time complexity and network performance.

Wavelength assignment algorithm in optical multicast networks with multi-wavelength conversion

Several structures of sharing wavelength converters (WCs) have been proposed for wavelength interchanging cross connects, including share-per-node (SPN), share-per-link (SPL), and share-per-wavelength (SPW). Among them, SPN has the highest sharing efficiency but at the expense of a complex switch fabric. In this paper, we investigate the Clos-type SPN and SPL structures and propose a novel structure, which is termed as C-SPW, to share fixed WCs based on the Clos network. The comparative simulations show that the sharing of fixed WCs in C-SPW can achieve similar drop probability (DP) with the sharing of tunable ones in Clos-type SPN under synchronous uniform traffic. The reason is that the fixed WCs in C-SPW can be accessible by all incoming packets, instead of a subset of them. Furthermore, an analytical model is derived to calculate the DP and converter utilization rate when C-SPW is operated in a synchronous mode under uniform traffic. The proposed analytical model fits well with the simulation results.

Placements of Shared Wavelength Converter Groups Inside a Cost-Effective Permuted Clos Network

In this paper, we propose a new composed buffer structure termed as wavelength converted broadcast-selective (WCBS). It copies each packet after conversion to all fiber delay lines, and makes the packet management flexible and efficient. By rearranging the packet waiting queue, the highest priority packet can leave the buffer without considering the existing queue, which makes preemption possible. Moreover, we construct an OPS node architecture employing the as-proposed WCBS buffer in output-buffered pattern. Some proper switching strategies are introduced to operate the switching node effectively. An analysis model is established to describe the packets in the node. The minimum packet loss performances under different traffic loads with different buffer parameters are calculated. The results are further verified by simulations. It implies that the WCBS buffer improves the performance of high priority packets effectively, especially when the high priority packet load is low.

A perturbation algorithm for embedding virtual data centers in multipath networks

Multi-wavelength conversion is a feasible technique to realize optical signal replication. However, there are little work to study the wavelength assignment and network performance in the optical network employing multi-wavelength conversion. In this paper, we investigate the wavelength assignment algorithm in optical networks with multi-wavelength conversion. To improve the resource utilization, multi-wavelength converters are shared at each node in the network. The main object of the proposed wavelength assignment algorithm is to maximize the acceptable destinations and minimize the wavelength conversion at the same time. Simulations are conducted to compute the blocking probability. It can be found that 25% multi-wavelength converters are sufficient to guarantee almost the same performance with full configuration of converters at each node, even when the traffic load is extremely high.

Deflection-compensated Birkhoff-von-Neumann switches

The Clos network is used for optical cross-connect and optical packet switch node, and has a large number of cross-points if the number of wavelengths per fiber is more than that of fibers per node. In this letter, we propose a cost-effective permuted Clos network to reduce the number of crosspoints. We further investigate the placement of fixed wavelength converter groups (WCGs) shared inside the permuted Clos network. On the contrary, a normal Clos network has to be equipped with expensive tunable WCGs. The simulation result implies that equipment of about 50% WCGs achieves similar performance as the full wavelength interchanging node does.

Allocation of wavelength selective and convertible cross connects in optical multicast networks

Network virtualization supports predicted network performance for applications by providing tenants with a virtual data center in multi-tenant data centers. The Hose model was recently extended for the virtual network abstraction and deployed in an Oktopus system, which offers the trade-off between cost and network performance. Embedding algorithms based on Oktopus model were well studied in the single-root tree topology. In this paper, we investigate congestion-aware allocation of virtual data centers in multipath networks. First, we formulate bandwidth constraints with linear programming, and provide a complete solution with exhaustive searching. Next, to reduce time complexity, we propose a perturbation algorithm to VM placement, which detects the bandwidth bottleneck of the current virtual machine placement and then adjusts the assignment to reduce congestion. The perturbation algorithm is compatible with both load-balanced and single-path routing algorithms. We compare the performance of the exhaustive searching algorithm and perturbation algorithm with load-balanced or single-path routing by simulations. The perturbation algorithm with load-balanced routing performs close to the exhaustive searching algorithm while significantly reduces the time complexity. Therefore, it offers a good tradeoff between time complexity and network performance.

Throughput performance of d-tree multicast capable optical cross-connect under synchronous traffic

Although the quasi-static scheduling based on Birkhoff-von-Neumann (BvN) decomposition can achieve high throughput with low operational complexity, its performance becomes less predictable when the input traffic is bursty. In this paper, we propose a deflection-compensated BvN (D-BvN) switch to enhance the performance. The D-BvN switch provides capacity guarantee for each virtual circuit (VC) by the BvN decomposition of average input traffic matrix, while copes with traffic burst by deflection. In particular, deflection scheme fully utilizes the spare capacity of starving VCs to deflect overflow traffic to other VCs and provide bandwidth for deflection traffic to re-access its desired VC. Analytical and simulation results show that it can achieve close to 100% throughput of offered load and the packet out-of-sequence probability is also negligible.