2008年4月2日 星期三

CoolStreaming/DONet: A Data-driven Overlay Network for Efficient Live Media Streaming

Xinyan Zhang, Jiangchuan Liu, Bo Li, Tak-Shing Peter Yum
CoolStreaming/DONet: A Data-driven Overlay Network for Efficient Live Media Streaming
In Proceedings of IEEE INFOCOM 2005

[Summary]
This paper presents Coolstreaming/DONet, which is a peer to peer media streaming system. For any peer to peer network system, member management, network structure, and content delivery management are critical issues to deal with. This paper performed a study to these issues. Furthermore, they evaluate this system through Planetlab environment.
DONet, which stands for data-driven overlay network, forms network structure by data availability. In previous peer to peer system, dedicated root node collects all information of nodes in this network and forms the tree, and data flow is also fixed, from roots to leaves. In DONet, however, network structure and data flow depends on if a data is available. Video data are sliced into segments, in this paper one segment mean one second. Each node keeps a buffer map that denotes data segment availability. Membership and partnership management is also a key issue in DONet. Each node keeps a partial member list called mCache. They use SCAM protocol to maintain and update membership list, and buffer map information. When a node enters this network, it contacts the server (or a specified node), gets a randomly selected deputy node, then gets partner candidates from deputy node’s mCache. And when a node wants to leave this network, it will issue a message. For the partnership management, each node periodically establishes new partnerships randomly from its mCache. An interesting scoring mechanism is used here, they calculate the average segment flow throughput, the higher the better. Better ones will be kept while this node is exploring new partners.
The scheduling algorithm is also an interesting issue in this paper. Since dynamic scheduling is a variation of NP-Hard problem. The author suggests a heuristic algorithm. This algorithm starts from knowing potential supplier for each segment. Since a segment with fewer suppliers has better chance that will miss the deadline, this algorithm starts from these segments, and scans these node for bandwidth. The one with the highest bandwidth and available time is selected.
This paper also presents the performance evaluation of this system. They use the PlanetLab platform to form the network, and perform tests on this system. They also termed CI, continuity index, which represents the smoothness of streaming, as a metric. Tests are performed in order to know CI and control overhead under different situation. Different streaming rate, partner number, and node dynamic period are considered. The performance of tree topology and DONet is also shown and compared.
The application of DONet is called Coolstreaming, which is a peer to peer streaming application. This study also reveals interesting facts. First, the current Internet has enough available bandwidth to support TV-quality streaming. The CI stays above 0.95 most of the time; this tells the smoothness of streaming. Second, as the nature of peer to peer system, the larger the network size, the better streaming quality obtained. The relationship of CI and user number is also shown.

[Comments]
Clarity:
This paper does well on clarity. The author stated some problems that every peer to peer system may meet, and come forward some solutions and observations. Started from previous work, the author discussed membership management, content management, transmission scheduling, implementation experience, and evaluation of this system. We can know the big picture of this system easily.
Correctness: The result and analysis of the evaluation seems to be reasonable. The function derived in this paper meets the result obtained.
Originality: The concept of DONet is quite novel. The flow of data sounds like energy balance in nature world. The new metric “CI” also tells the fact of streaming experience.
Significance: Due to that it is quite complete study; I think this paper is significant. The concept of DONet impacts the design of p2p network. The experiences of peer to peer streaming application is also valuable.

2008年3月12日 星期三

End-to-End Internet Packet Dynamics

Vern Paxson
End-to-end internet packet dynamics
ACM/IEEE Transactions on Networking, 7(3):277-292, June 1999

[Summary]
This paper mainly focuses on the TCP traffic analysis between sites, while TCP is widely used in real world. They made an analysis based on the measurement of TCP transfer traffic, and network behavior. The measurement is obtained by sending and receiving 100 Kbyte TCP transfer between 2 of 35 sites, which are separately located in 9 countries in US and Europe, and they were taken in 1994 and 1995, and these two set of data are called N1 and N2. N2 has enlarged window size ensuring TCP connection throughput. In this paper, they listed and analyzed some prevalent network events like packet loss, network pathologies, detection of bandwidth bottleneck in the path, and packet delay.
Packet loss is the important thing that every networking people care about. In this paper, the author wants to find out a general pattern of packet lost. The author claims that acks are better for determining packet loss rate because of transmission rate of acks is stable. It’s also claimed that packet loss is somehow correlated. That is, if a packet is lost, the next packet has good chance to lose.
Packet delay is also a hot topic. Due to asymmetric path of packets, delay is calculated in one direction, called one-way transit time (OTT). And this would be modeled to a certain distribution. Furthermore, the author tries to find out available bandwidth.
Network pathology, meaning unusual or unexpected packet delivery, is crucial to advance analysis. This paper talked about out-of-order delivery, packet replication, and packet corruption. For out-of-order delivery, which has little impact on TCP performance, effects TCP "duplicate acks" that determines if the packet needs re-transmission. They concluded that out-of-order delivery is site independent, and asymmetric. Packet replication, which is an unnecessary retransmission event that sender mistake the packet was not properly received, is also site independent. Packet corruption occurs when the transmission channel has noise or some error while going through routers.
In this paper, bottleneck bandwidth is also discussed. Bottleneck bandwidth means the upper limit on how quickly the network can deliver the sender’s data to the receiver. The bottleneck estimation technique used is based on the idea of “packet pair”. The sender sends consecutive packets and measures the time interval of these packets at sending and receiving. But somehow this method suffers from several potential problems like out-of-order delivery. Motivated by the shortcomings of “packet pair”, packet bunches, instead of pairs, are used in bottleneck estimation.

[Comments]
Clarity: I think this paper is a good at clarity. The author made a topic he wants to discuss at first. And the reason of this topic is explained. The measurement method and result analysis is illustrated. Several graphs and tables are used to express the measurement result, and the author explained these figures and tables well.
Correctness: The author shows some graphs of measurement result that is close to what I think it would be and the phenomenon behind these measurements are discussed. However, several problems are stated but still not having a proper answer yet.
Originality: This paper measured TCP network traffic and brought up some new view to these measurement.
Significance: I will say this paper is significant. For all of networking people, TCP traffic and packet dynamics analysis is crucial.