Internetworked Systems Lab


Meeting No 29

Date, Place: Thursday, 19/11/09, 14:00, InterSys Lab
Presenters: Sotirios-Angelos Lenas, Nikolaos Bezirgiannidis
Topic: The DINET Experiment

Presentation Abstract: In October and November of 2008, Jet Propulsion Laboratory under contract to Nasa installed and tested essential elements of Delay/Disruption Tolerant Networking technology on the Deep Impact spacecraft and on nine other nodes at JPL. In this talk we will present you the basic elements of this experiment, called the Deep Impact Network Experiment (DINET). We will also point out some qualitative results, the DTN-related anomalies which occurred during the testing period and finally the significance, both technical and strategic, of this experiment.


Meeting No 28

Date, Place: Thursday, 05/11/09, 14:30,  InterSys Lab
Presenter: Arsakian Armen
Topic: Characterizing Internet traffic: When the point of view matters

Presentation Abstract: Internet is constantly evolving mainly due to the introduction of new applications on the endpoints while ,at the same time, link capacities increase. Trying to understand the interaction of TCP/IP stacks with the core of Internet remains a challenge. During the presentation I will give examples of Poisson and Bursty traffic from real measurements on LAN using traffic generators. Also, I will explain the importance of sample size when we examine statistical properties. Finally I will demonstrate that the PoP location counts, when we estimate distributions like interarrival times with respect to access links. The simulations performed show how non-responsive traffic interacts with the path available capacity in terms of changes to interpacket times.


Meeting No 27

Date, Place: Thursday, 22/10/09, 15:00, InterSys Lab
Presenter: Ioannis Komnios
Topic: Performance evaluation of networks that utilize reliable link and transport layer protocols

Presentation Abstract: In this talk, we extensively study reliability mechanisms in both transport and link layers, in wireless and satellite networks, and we underline the advantages of each method. We study the overhead that entails the implementation of reliability mechanisms in two distinct network layers and the correspondent performance increase. Initially, through a set of reference scenarios, we study the impact of reliable transport and link layer protocols in terrestrial wireless networks, which suffer from errors or congestion. Next, we continue with a set of innovation scenarios, where we study the performance of reliable protocols in satellite networks, which are characterized by high propagation delay and error rate values. Moreover, we study the impact that the number of intermediate nodes between the sender and the receiver has on system performance, when propagation delay values are high.


Meeting No 26

Date, Place: Thursday, 15/10/09, 14:30, InterSys Lab
Presenter: Giorgos Papastergiou
Topic: Erasure-Coding Based Routing for Opportunistic Networks / Application of Long Erasure Codes at Transport Layer

Presentation Abstract: Routing in Delay Tolerant Networks (DTN) with unpredictable node mobility is a challenging problem because disconnections are prevalent and lack of knowledge about network dynamics hinders good decision making. Current approaches are primarily based on redundant transmissions. They have either high overhead due to excessive transmissions or long delays due to the possibility of making wrong choices when forwarding a few redundant copies. In this presentation, i will present a novel forwarding algorithm based on the idea of erasure codes. Erasure coding allows use of a large number of relays while maintaining a constant overhead, which results in fewer cases of long delays. Based on the concept of using erasure-coding instead of redundant transmissions, i will present several research issues regarding the application of long erasure codes at the Transport Layer, focusing mainly on Deep-Space Transport Protocol, whose retransmission scheme is currently based only on redundant transmissions.


Meeting No 25

Date, Place: Thursday, 08/10/09, 15:00, InterSys Lab
Presenter: Efthymios Koutsogiannis
Topic: A DTN Testbed Architecture Presentation

Presentation Abstract: As the space communications infrastructure of the near future is yet to be standardized, many protocols have been proposed to cope with issues such as long propagation delays, high packet error rates and intermittent connectivity. In that context, we have designed a testbed, which incorporates Space conditions, mission scenarios and Space communication requirements, rendering the extensive evaluation and design of space protocols feasible and efficient.


Meeting No 24

Date, Place: Thursday, 01/10/09, 15:00, InterSys Lab
Presenter: Christos Liaskos
Topic: Balancing Implementation Cost and Performance: An Overview of Research on Wireless Broadcast-based Systems.

Presentation Abstract: Data dissemination through broadcasting has been gaining popularity over the past years. Many modern wireless telecommunication systems suffer from scalability, population coverage and QoS problems. Overcoming these shortages through the traditional pull-based logic is rather costly, as it entails the installment over pretty powerful hardware at the server side. Data broadcasting on the other hand has been adopted by the IT industry as a cheap alternative, either in the form of multicasting/broadcasting-enabled Internet routers and related protocols or push-based services via cellular networks. In the context of this presentation the application boundaries between pull and push architectures will be discussed, and the theoretical broadcast problem will be defined. Related outstanding analytical and heuristic approaches, as well as the research of the Computer Architecture and Communications Lab on the field will be presented, introducing adaptive push-based systems, analysis-based broadcast scheduling approaches and top-performing cost-efficient scheduling algorithms. Finally, the prospect and advantages of uniting pull and push approaches into a hybrid system will be explored.


Meeting No 23

Date, Place: Thursday, 24/09/09, 15:00, InterSys Lab
Presenter: Sotirios-Angelos Lenas
Topic: Comparison between dynamic and static routing on networks with constantly changing connectivity maps

Presentation Abstract: In this talk I will present the differences in performance on networks that utilize constantly changing connectivity maps, when static and dynamic routing is applied. Especially on networks with extreme conditions, such as long propagation delays and lack of connectivity, the selection of the most efficient routing method, pose difficulties. The simulation results that i'll present, show that static routing outperforms dynamic routing for high propagation delay values, although it is affected more by long periods of lack of connectivity and connection errors.


Meeting No 22

Date, Place: Thursday, 17/09/09, 14:30, InterSys Lab
Presenter: Stylianos Dimitriou
Topic: Effective Buffer and Storage Management in DTN Nodes

Presentation Abstract: Current wired networks have been developed on the basis of the AIMD principle, which offers increased performance and fairness. Nevertheless, there is a vast spectrum of networks, from deep space to wireless sensor networks, where TCP fails to operate, as it results in frequent timeouts and intermittent or no connectivity at all. The DTN (Delay/Disruption Tolerant Networking) architecture and particularly the Bundle protocol, offers solutions in networks where low connectivity, big or variable delays and increased error rates are prevalent. Using persistent storage, a DTN node is able to overcome such limitations; the data can be transmitted several times until it is successfully received by the next hop. However, since the conditions in a DTN network might vary, from time to time we may experience high connectivity and low delays. In such cases, data does not need to be transferred to persistent storage, as it will be probably be transmitted shortly and this transfer from buffer to persistent storage and back to buffer, will delay the transmission. In this paper we propose a system which implements a mechanism that aims to minimize packet transfers between buffers and persistent storage and hence accelerate transmissions.


Meeting No 21

Date, Place: Thursday, 03/09/09, 15:00, InterSys Lab

1st Presenter: Sotiris Diamantopoulos
Topic: Evaluation of Dynamic DTN Routing Protocols in Space Environments

Presentation Abstract: As the number of space elements increases, routing becomes an issue of great interest. The majority of routing schemes that have been proposed till now employ a fairly static design and only recently more sophisticated protocols have been introduced. In this study, we evaluate some of the most prominent routing protocols for Delay-Tolerant Networks, Epidemic, PRoPHET and Contact Graph Routing, in space environment. Using COMNET Lab’s DTN testbed, we show that for increasing delay, Contact Graph Routing significantly outperforms the other two routing schemes.


2nd Presenter: Efthymios Koutsogiannis
Topic: Evaluation of CCSDS File Delivery Protocol over Delay Tolerant Networks

Presentation Abstract: Space communications enter a new era where a multihop architecture is required and an increasing number of alternative communication paths may be available. Current space applications, such as CFDP for file transfer, rely on static routing and can not efficiently perform in these challenging environments. The most promising solution is the emerging Delay Tolerant Networking (DTN) Architecture and the accompanying Bundle Protocol which allow for reliable store-and-forward message switching, custody transfer and dynamic routing. However, a space-oriented file transfer protocol for DTN is required. In this work, we integrate CFDP with ION (a Bundle Protocol implementation produced by the Jet Propulsion Laboratory) and evaluate its performance in a DTN testbed.


3rd Presenter:
Ioannis Komnios
Topic: Comparison of Static and Dynamic Routing in Networks that utilize Constantly Changing Connectivity Maps

Presentation Abstract: Routing in computer networks has always attracted research interest; throughout the years different solutions have been proposed, from simple static routing to more intelligent routing algorithms. Extreme network conditions, like long propagation delays and lack of connectivity, pose difficulties for the selection of the most efficient routing method. In this paper, we study the performance of networks that utilize constantly changing connectivity maps, when static and dynamic routing is applied. Our simulation results show that static routing outperforms dynamic routing for high propagation delay values, although it is affected more by long periods of lack of connectivity.


Meeting No 20

Date, Place: Thursday, 13/11/08, 15:00, InterSys Lab
Topic: Discussion on research topics of the group

Meeting No 19

Date, Place: Wednesday 06/02/07, 10:00, InterSys Lab
Presenter: Christos Samaras
Topic: DTTP: a Delay-Tolerant Transport Protocol for Space Internetworks

Presentation Abstract: We propose Delay-Tolerant Transport Protocol (DTTP) to address reliable data transfer in stressed network environments, such as space communications. Since existing TCP mechanisms do not work well (or at all) for such networks, new transport schemes are required. Intermittent connectivity in space environments calls for new transport approaches that smoothly adapt to the special networking conditions. DTTP is primarily a transport layer protocol and satisfies the inherent architecture requirements of Delay Tolerant Networking (DTN) in the absence of IP network infrastructure. It allows for reliable, efficient data transfer offering a number of application-oriented transmission strategies. Otherwise, when an IP architecture exists, DTTP operates as a stand-alone transport entity which interfaces with IP directly. We introduce the protocol's properties and functionality that enable its deployment in challenged net-works. We conduct simulations that demonstrate the protocol's efficiency in scenarios with: (i) long propagation delays, (ii) minimum to relatively high packet error-rate, and (iii) intermittent connectivity.


Meeting No 18

Date, Place: Friday, 14/12/07, 10:00, InterSys Lab
Presenter: Giorgos Papastergiou
Topic: Deep-Space Transport Protocol (DS-TP)

Presentation Abstract: We present Deep-Space Transport Protocol (DSTP), a new reliable protocol for deep-space communication links. DS-TP’s main advantage is its ability to complete file transfers faster than conventional TCP, SCPS-TP and Saratoga. Therefore, missions with small connectivity time are greatly favored. Deep space communication links are characterized by long propagation delays, high BERs, intermittent connectivity (i.e., blackouts) and bandwidth asymmetries. Common approaches to deal with the above unique characteristics are: rate-based, open-loop protocols to deal with huge propagation delays; regular retransmissions to deal with high BERs; transmission suspension to deal with blackouts; SNACKs to deal with bandwidth asymmetries. We adopt some of the above approaches, namely, the open-loop, rate-based transmission and the SNACKs and focus on the optimization of the rest, namely, the retransmission strategy of the transport protocol to deal either with high BERs or with blackouts. More precisely, DS-TP includes the Double Automatic Retransmission (DAR) technique. DAR sends each packet twice, importing some intentional delay (Rd) between the original transmission and the retransmission. Therefore, in the presence of communication gaps (i.e., errors or blackouts), corrupted packets will eventually be replaced by the same correct packets that arrive with delay Rd. Rd, however, is much smaller than the traditional TCP-RTO value. Our theoretical performance evaluation results reveal that DS-TP presents high potential for deployability. In particular, we show that for PER=50%, DS-TP completes a file transfer in half time of a conventional protocol.


Meeting No 17

Date, Place: Wednesday, 05/12/07, 10:00, InterSys Lab
Presenter: Ioannis Komnios
Topic: Routing in Delay Tolerant Networks

Presentation Abstract: Routing in Delay Tolerant Networks (DTNs) is of critical importance since routing protocols in DTN need to increase the probability of successful message delivery using highly restricted information. There are three different points in DTNs where routing can take place: Source Routing (the route is determined in the source node), Per Hop Routing (every hop determines every next hop along with the routing path) and Per Contact Routing (routing table is recalculated every time a new contact is found). In this presentation we analyze four different routing methods in Delay Tolerant Networks. In Epidemic Routing all messages are randomly replicated until all nodes have a copy of every message. Delay Tolerant Link State Routing is similar to Classic Link State Routing, although now even connections that are temporarily unavailable are taken into consideration when calculating the route. Knowledge Oracles are used to comprehend the tradeoff between knowledge and performance. MobySpace uses Euclidean Space as a tool to help nodes make routing decisions.


Meeting No 16

Date, Place: Wendsday, 28/11/07, 10:00, InterSys Lab
Presenter: Theodoros Amanatidis
Topic: Delay Tolerant Networking Architecture

Presentation Abstract: Delay and Disruption Tolerant Networking Architecture (DTN) embraces the concepts of occasionally connected networks that may suffer from frequent partitions and that may be comprised of more than one divergent set of protocols. The basis for this architecture lies with that of the Interplanetary Internet, which focused primarily on the issue of deep space communication in high-delay environments. DTN architecture can be utilized in various operational environments, including those subject to disruption and disconnection and those with high-delay; the case of deep space is one specialized example of these, and is being pursued as a specialization of this architecture. Other networks to which this architecture applies include sensor-based networks, terrestrial wireless networks, satellite networks, and underwater acoustic networks. DTN architecture consists of an end-to-end message-oriented overlay called the "bundle layer" that exists at a layer above the transport (or other) layers of the networks on which it is hosted and below applications. Bundle layer includes a hop-by-hop transfer of reliable delivery responsibility and optional end-to-end acknowledgement.


Meeting No 15

Date, Place: Friday, 11/05/07, 15:30, InterSys Lab
Presenter: Ageliki Tsiolaridou
Topic: On the Miscollaboration of Congestion Control Mechanisms at the Transport and the Network Layers

Presentation Abstract: Many sophisticated mechanisms have been implemented at both the transport and the network layers in order to estimate network conditions and avoid overload of network links. We evaluate the congestion control and avoidance scheme from a cross-layer perspective. We conduct simulations to investigate issues of miscollaboration between the network and transport layers that have negative impact on application performance. We identify several such scenarios; we discuss the reasoning and the extent of miscollaboration, and the appropriate measures to resolve the problem. We conclude that the common practice for evaluating research proposals lacks the appropriate experimental diversity, granularity or perspective and hence new mechanisms may favor a particular set of targeted protocols but damage some others. As a result, a portion of Internet traffic flows is frequently favored at the cost of co-existing traffic flows.

Meeting No 14

Date, Place: Thursday, 08/05/07, 16:00, InterSys Lab
Presenter: Ioannis Psaras
Topic: The TCP Minimum RTO Revisited

Presentation Abstract: We re-examine the two reasons for the conservative 1-second Minimum TCP-RTO to protect against spurious timeouts: i) the OS clock granularity and ii) the Delayed ACKs. We find that reason (i) is canceled in modern OSs; we carefully design a mechanism to deal with reason (ii). Simulation results show that in next generation's high-speed, wireless-access networks, TCP-RTO should not be limited by a fixed, conservative lower bound. For further information on the topic, check the NETWORKING 2007 paper.


Meeting No 13

Date, Place: Friday, 19/01/07, 13:30, InterSys Lab
Presenters: Ageliki Tsiolaridou, Lefteris Mamatas, Eleni Kamateri
Topic: Technologies Related to Router Queues

Presentation Abstract: Router buffers are crucial elements of packet networks. They accommodate transient bursts in traffic, minimizing drop events; furthermore, they keep a reverse of packets so that the link does not go idle. Internet communication should be supported by optimum sized router buffers that can be exploited from mechanisms performing traffic shaping or/and supporting end-to-end congestion control. We are going to discuss about the following technologies: (i) The objectives of buffer sizing in routers, rules that have been recently proposed and issues that can have major impact in the buffer sizing process. In the last years, several papers have proposed new rules that determine the appropriate buffer sizing in the Internet routers. (ii) A new mechanism that supports delay-sensitive applications through alternative packet scheduling, called Traffic Sensitive QoS (TSQ). TSQ provides better delay performance for delay sensitive applications and higher throughput for throughput sensitive applications. (iii) A simple, low complexity protocol, called Variable-structure congestion Control Protocol (VCP) that leverages only the existing two ECN bits for network congestion feedback and yet achieves comparable performance to XCP.

Meeting No 12

Date, Place: Thursday, 21/12/06, 15:30, InterSys Lab
Presenters: Panagiotis Papadimitriou, Giorgos Papastergiou
Topic: Transport-layer Perspectives for the Interplanetary Internet

Presentation Abstract: Interplanetary Internet (IPN) constitutes a highly challenging environment and is envisioned to allow a plethora of communication services. IPN should be able to encompass both terrestrial and interplanetary links, enforcing communications across disparate environments. Initially, we investigate the suitability of existing transport-layer solutions to efficiently deliver audio and visual information from local environments to earth stations. As a viable alternative, we provide an overview of RCP-Planet: a protocol that deploys a newly developed end-to-end rate control scheme based on rate probing. RCP-Planet addresses most IPN challenges: long propagation delays, increased bit error rates, bandwidth asymmetry and blackout effects. We demonstrate several simulation results to evaluate the protocol's performance. RCP-Planet appears to outperform the most remarkable transport-layer solutions for time-sensitive traffic, such as TFRC, RAP and SCTP. Furthermore, we briefly refer to the architecture and the elements of Delay-Tolerant Networks, focusing on the role of the bundle layer that binds operationally the application with the lower layers, in the context of a future space/ground protocol stack. In the sequel, we present the Space Communications Protocol Standards-Transport Protocol (SCPS-TP), a set of extensions to TCP for deep-space networks. We further analyze the throughput performance and transmission behavior for the rate-based and window based congestion control mechanisms in a small satellite environment. This analysis reveals that the traffic shaping mechanism of rate-based transmission is more effective than the bursting flow of a window-based mechanism in error-prone space environments with long link-delays.

Meeting No 11

Date, Place: Thursday, 14/12/06, 15:00, InterSys Lab
Presenters: Ioannis Psaras, Akis Kontonikolas
Topic: Extending Internet into Space - On Transport Layer Issues For InterPlaNetary Networks

Presentation Abstract: The primary objective of space exploration missions is to gather information for neighbor planets, the space and the universe. A fundamental requirement of such missions is efficient and reliable communication between scientists (on Earth) and spacecrafts or astronauts (in space). Therefore, communication problems, of any kind, need to be efficiently solved. Communication scientists and engineers tend to correlate space networks with the Internet. Differences, however, between a Planetary and an InterPlaNetary Network are not yet well addressed. In this presentation we will, initially, talk about the major challenges in Transport Layer Networking in Space. We will give examples, present related work, criticize proposed solutions and discuss design guidelines for an efficient and integrated Transport Protocol for Space Communications. We will then present, in detail, one such proposed protocol, namely TP-Planet, which is said to efficiently deal with InterPlaNetary Networks.

Meeting No 10

Date, Place: Thursday 06/04/06, 15:00, InterSys Lab
Presenter: Ioannis Psaras
Topic: WB-RTO: A Window-Based Retransmission Timeout for TCP

Presentation Abstract: Based on recent research, we have observed that TCP timers do not always adjust to the level of flow contention. Several flows may calculate the same timeout, leading to congestion events due to simultaneous retransmissions. We present a new timeout algorithm for TCP, based on the observation that TCP-RTO should not be solely based on RTT estimations. We argue that the design principles of the current timeout algorithm may lead to flow synchronization, unnecessary retransmission effort and unfair resource allocation. The Window-Based Retransmission Timeout (WB-RTO) exhibits two major properties: (i) it cancels retransmission synchronization which dominates when resource demand exceeds by far resource supply and (ii) reschedules flows on the basis of their contribution to congestion. Our simulation results match our design goal. For further information on the topic, check the GLOBECOM 2006 paper or an extended Journal version (Elsevier COMNET 2006).


Meeting No 9

Date, Place: Thursday, 23/03/06, 18:00, InterSys Lab
Presenter: Lefteris Mamatas
Topic: Energy Issues for Next Generation Networking

Meeting No 8

Date, Place: Thursday, 16/03/06, 15:00, InterSys Lab
Presenter: Tobias Harks
Topic: User Strategies in Networks Implementing Congestion Pricing

Presentation Abstract: In the context of a network implementing congestion pricing, we focus on user strategies to determine their willingness to pay parameter. We argue that users downloading a file in fixed time, or users running a multimedia application will have other strategies to decide on their payment than simply to maximize their surplus measured by a concave utility function minus cost. Instead, we formulate a download task as an optimal control problem and account for dynamic changes of the state of congestion by using (online) model predictive control techniques. Finally, we develop online control strategies in order to automatically adapt the willingness to pay parameter. For further information on the subject check the CCNC06 paper.


Meeting No 7

Date, Place: Thursday, 26/01/06, 18:00, InterSys Lab
Presenter: Christos Samaras
Topic: Game Theory and Computer Networks: a useful combination?

Presentation Abstract: In a complex environment like the Internet and in computer networks generally, many problems arise when greedy users (or transport entities/protocols) transmit data at high rates trying to gain more than equal share of channel bandwidth even when congestion is present or when users act irrationally by putting a lot of effort to achieve a better throughput though their payoff actually decreases. The aforementioned scheme calls for network rules and actions that will motivate the behavior of users. Moreover, a set of transmitting strategies needs to be defined and become available to users, so that they (the users) can make a rational strategy choice in a distributed environment like the Internet where conditions often change rapidly and proper adaptation becomes necessary. We will present Game Theory, a general theory of rational behavior. Various aspects of Game Theory (such as Utility Function and Nash Equilibrium) will be discussed. Can Game Theory become a useful tool for regulating data transmission in computer networks? Can it serve as a suitable method for developing a model of Internet congestion control and provide answers to open problems in the Internet?


Meeting No 6

Presenter: Ageliki Tsioliaridou
Topic: Network and Protocol Mechanisms: How well do they collaborate?


Meeting No 5

Date, Place: Thursday, 15/12/05, 18:00, InterSys Lab
Presenter: Ioannis Psaras
Topic: CA-RTO: A Contention-Adaptive Retransmission Timeout for TCP

Presentation Abstract: Based on recent research, I have observed that TCP timers, based solely on RTT estimations and measurements, cannot capture with precision the level of flow contention. I noticed that high level of flow contention may stabilize RTT variation, minimize the deviation and, in turn, shorten the timeout value. I prove that this behavior is undesirable indeed, since it leads to unfair resource utilization. I propose CA-RTO
(Contention – Adaptive Retransmission Timeout), an algorithm that incorporates a contention parameter and a randomization technique into the Retransmission Timeout algorithm of TCP. Finally, I verify the strength of my proposal based on simulation results and report significant improvement in fairness, great reduction of retransmitted packets and slight improvements in application goodput. For further information on the topic, refer to the ICCCN05 paper.


Meeting No 4

Date, Place: Thursday, 08/12/05, 18:00, InterSys Lab
Presenter: Panagiotis Papadimitriou
Topic: Transport Layer Mechanisms for Real-Time Application QoS

Presentation Abstract: Internet is evolving into a universal communications network, hosting several types of traffic including traditional data, voice and video. Providing Quality of Service (QoS) is challenging concerning the heterogeneity of the Internet, as well as the real-time applications' stringent QoS requirements. In this context, I discuss whether the current best-effort technology of the Internet stands up to the high quality standards of time-sensitive applications. Exploiting a new metric for the evaluation of real-time application performance, I reach several conclusions on the performance of streaming media delivery and VoIP quality from the perspective of transport protocol support and efficiency. Furthermore, I present a framework for adaptive video streaming over UDP relying on a basic and efficient end-to-end congestion control mechanism. The proposed architecture employs a new transport protocol, namely Scalable Streaming Video Protocol (SSVP), which is able to interact with new and existing rate-adaptive video streaming applications based on receiver side congestion feedback. SSVP attempts to maximize the performance of streaming video delivery with concern to friendliness with interfering traffic. Finally, I refer to the challenges of real-time streaming over satellite IP networks, since satellite links commonly exhibit long propagation delays and increased error rates, which impair TCP performance. In this context, I quantify the effects of satellite links on TCP efficiency and streaming video delivery. In addition, I study the supportive role of Selective Acknowledgments (SACK), address the associated slow-start issues and evaluate the impact of delayed acknowledgments.


Meeting No 3

Date, Place: Thursday, 01/12/05, 18:00, InterSys Lab
Presenter: Lefteris Mamatas
Topic: Energy Issues for Next Generation Networking

Meeting No 2

Presenter: Panagiotis Papadimitriou
Topic: Diffserv


Meeting No 1

Presenter: Panagiotis Papadimitriou
Topic: Linux Traffic Control (TCNG)