Multi-tier platforms are now widely used to build Internet application servers. These platforms are usually replicated on clusters of machines to achieve scalability and availability. This paper presents the main principle to wrap legacy software pieces in components in order to provide a uniform management interface, thus allowing the implementation of management applications. Management applications are used to deploy distributed applications and to autonomously reconfigure them as required.
Transmission Control Protocol (TCP) is one of the main transmission protocols used in the Internet. It has been observed that opening parallel TCP sessions might be of interest for a user in order to increase his overall average throughput. In this paper, the researchers suggests to charge users per TCP session, and they investigate the resulting game in a homogeneous context: how many sessions should each user open? Given the discrete (and even finite) space of strategies, they also propose to implement a probabilistic adaptation algorithm, analyze its theoretical properties and provide numerical illustrations.
The huge success of the Internet and its transformation into an important commercial infrastructure is in fact a double edge weapon. It is now necessary for Internet designers to consider new consumer expectations in term of performance, services and bandwidth demand. Satellite networks and Asymmetric Digital Subscriber Lines (ADSL) are promising technologies since they offer significant bandwidth increases over download paths of existing Internet traffics. This paper from French National Institute for Research in Computer Science and Control presents a class-based scheduler called ACQ for bandwidth sharing between Data and ACK flows in an asymmetric bandwidth environment.
This research paper studies pricing of differentiated services and its impact on the choice of service priority at equilibrium. The paper considers both TCP connections as well as non controlled (real time) connections. The performance measures (such as throughput and loss rates) are determined according to the operational parameters of a RED buffer management. The latter is assumed to be able to give differentiated services to the applications according to their choice of service class.
The performance of competing connections that share a bottleneck link is studied in this paper. The paper assumes that data transfer uses TCP congestion control protocol and that real-time traffic uses some TCP- friendly transport protocol that satisfies the same square-root formula for throughput. A Stackelberg game is analyzed where users choose the price they are willing to pay at a RED buffer in order to discriminate service. This simplification allows obtaining explicit formulae for the equilibrium priority choices of users.
Bandwidth sharing between multiple TCP connections has been studied under the assumption that the windows of the different connections vary in a synchronized manner. This synchronization is a main result of the deployment of Drop Tail buffers in network routers. The deployment of active queue management techniques such as RED will alleviate this problem of synchronization. This white paper explains a mathematical model to study how the bottleneck bandwidth will be shared if TCP windows are not synchronized. This permits to evaluate the improvement in fairness and utilization brought by the deployment of active buffers.
TCP congestion control is often analyzed using linear-increase multiplicative-decrease models for window variation. This paper develops a model for TCP that accounts for both sub linearity and limitation of window increase. Sub linear window growth is observed when the round-trip time of the connection increases with the window size. The limitation is due to the window advertised by the receiver. First, the paper derives the required conditions for the stability of the model. Then, it writes the Kolmogorov equation under Markovian assumptions. The model is solved analytically for some particular cases.
The TCP protocol performance degrades over wireless links since it uses the loss of packets as an indication of network congestion, which are characterized by a high bit error rate. Different solutions have been proposed to improve the performance of TCP over wireless links, the most promising one being the use of a hybrid model at the link-level combining FEC, ARQ-SR(Automatic Repeat Request with Selective Repeat), and an in-order delivery of packets to IP. The drawback of FEC is that it consumes some extra bandwidth to transmit the redundant information.
Accessing the Internet via asymmetric paths is becoming common with the introduction of satellite and cable networks. Users download data from the Internet via a high speed link and send requests and acknowledgements (ACK) via a slow reverse channel. ACK filtering has been proposed as a technique to alleviate the congestion on the reverse path of a TCP connection. In the literature the case of a one-ACK per connection at a time in the buffer at the input of a slow channel has been studied. This paper shows that this is too aggressive for short transfers.
This white paper illustrates a simple scheduler called SBQ (Service-Based Queuing) to share the bandwidth fairly between unicast and multicast flows according to a new definition of fairness referred as the inter-service fairness and describes a new active queue management mechanism called MFQ (Multicast Fair Queuing) to fairly share the allowed multicast bandwidth among competing flows in the multicast queue. The reproduction results obtained for very heterogeneous sources and links characteristics suggest that, on one hand, SBQ attains the expected aggregated bandwidth sharing among unicast and multicast service and on the other hand the multicast flows remain TCP-friendly.