Zhenghu Gong

OpenSAN: a software-defined satellite network architecture

In recent years, with the rapid development of satellite technology including On Board Processing (OBP) and Inter Satellite Link (ISL), satellite network devices such as space IP routers have been experimentally carried in space. However, there are many difficulties to build a future satellite network with current terrestrial Internet technologies due to the distinguished space features, such as the severely limited resources, remote hardware/software upgrade in space. In this paper, we propose OpenSAN, a novel architecture of software-defined satellite network. By decoupling the data plane and control plane, OpenSAN provides satellite network with high efficiency, fine-grained control, as well as flexibility to support future advanced network technology. Moreover, we also discuss some practical challenges in the deployment of OpenSAN.

FlyCast: Free-Space Optics Accelerating Multicast Communications in Physical Layer

In this paper, we propose FlyCast, an architecture using the physical layer of free-space optics (FSO) to accelerate multicast communication. FlyCast leverages off-the-shelf devices (e.g. switchable mirror, beam splitter) to physically split the FSO beam to multi receivers on demand, which enables to build dynamical multicast trees in physical layer and accelerates multicast communications. We demonstrate the feasibility of FlyCast through our theoretical analysis and the proof-of-concept prototype.

Routing Algorithm Based on Ant Colony Optimization for DTN Congestion Control

Due to the limitation of DTN resource and intermittent connection, its challenging to design a mechanism which achieves high message delivery successful ratio and control congestion to satisfy the demand. In this paper, a novel DTN congestion control routing mechanism, Ant Colony Optimization(ACO) for DTN congestion control algorithm(ADC) is proposed. In the message delivery, this mechanism can actively leverage pheromone and heuristic in ACO to evaluate the selection of transfer node. Pheromone can guarantee the message successful delivery ratio and heuristic prevent network congestion in real time. The evaluation of this congestion control routing algorithm shows the evident improvement of some parameters in network circumstance.

SARM: A Congestion Control Algorithm for DTN

When congestion happens at the node of DTN, the regular approach is to delete the old message of these nodes in order to make memory space for the new arrival message to remove congestion. A new method for the selection of message to delete, congestion control routing algorithm based on Simulated Annealing and regional movement (SARM) are proposed in this paper. This mechanism is under the premise that node movement has the regional characteristic. The probability of node collision and transferring nodes in the message delivery are taken into account comprehensively. We quantitatively evaluate the successful probability of different messages delivered by different nodes. When node congestion happens, we choose the message with lower probability of successful delivered by two encounter nodes to delete using SARM. The experiment shows that the parameters of network are significantly improved under this congestion control routing algorithm.

Following Routing: An Active Congestion Control Approach for Delay-Tolerant Networks

Due to the limited resources and intermittent connectivity of Delay-Tolerant Networks (DTNs), it is a challenging task to design an efficient congestion control method to meet the application requirements. In this paper, a new DTN congestion control method called following routing(FR) is proposed. In the method, when the transit node is about to be congested, current node will follow the route of the transit node to avoid packet drop-off due to memory shortage, and finally avoid congestion. We use the Markov process to determine the timing enabled to use FR. The new congestion control method is evaluated through simulation and the experimental results have shown that network performance can be improved significantly.

HERO: A Hybrid Electrical and Optical Multicast for Accelerating High-Performance Data Center Applications

This paper presents HERO, a system for accelerating high performance data center applications by integrating hybrid electrical and optical multicast. We built a prototype and developed a flow-level simulator to evaluate the performance of HERO. Experiment and simulation results show that HERO reduces the average MFCT by ~32% and ~28% compared to OCS and EPS multicast, respectively.

A Novel Resource Allocation Heuristic for a Power Constrained Data Center

Nowadays Internet-scale data centers are facing a severe issue, huge energy consumption, which not only increases operational cost, but also causes environment problem. For data center providers, putting power constraint on a data center may be a feasible way to keep energy consumption within an acceptable range. Hence it is commonly believed that, for a given power budget, maximizing the performance becomes an important goal for a data center, which can be achieved with well-designed resource allocation algorithm. In this paper, we discuss the resource allocation problem in a power-constrained data center, where the power consumption of whole cluster should not exceed a threshold that is set by data center provider. Then, we propose a novel resource allocation heuristic based on preference and fair allocation policy to this problem, and present further discussion and conclusion at the end.

CRASP: congestion control routing algorithm against selfish behavior based on pigeonhole principle in DTN

In DTN, a small amount of selfish nodes occupy the storage of specific key nodes with stronger transfer capacity in hot spot region which greatly consumes the limited network storage resource. This causes congestion and reduces the successful delivery ratio of messages from other nodes. This paper adopts pigeonhole principle and power in combinatorics to define nodes responsibility range of message transfer for other nodes. It is helpful for the nodes to selectively accept messages which are required to be transferred and deliver them. The algorithm should restrict the selfish behavior of selfish nodes, reduce network congestion and guarantee the reasonable share of storage resource by all nodes according to the communication frequent degree. The experiment proves that congestion control routing algorithm against node selfish behavior based on pigeonhole principle (CRASP) proposed in this paper excellently achieves the above objectives.

Routing Algorithm Based on Nash Equilibrium against Malicious Attacks for DTN Congestion Control

In Delay-Tolerant Network(DTN), certain malicious node might generate congestion in attack to reduce the overall performance of the whole network, especially the target of message successful delivery ratio. In this paper, a novel Nash equilibrium based congestion control routing algorithm with the function of security defense (NESD) is proposed. In the process of message delivery, node can use Nash equilibrium to compute the largest proportion of transfer messages occupancy to node memory capacity. This mechanism constrains the attack from malicious node and guarantees the message transfer of regular node. This congestion control routing algorithm for security defense is evaluated by experiment. It is important application in the field of homeland defense. The results show that related key parameters are significantly improved in DTN scenario.