Łukasz
Budzisz
|
Hamilton Institute National University of Ireland (NUI) Maynooth Maynooth, Kildare Ireland Tel.:
+353
1 708-6273 |
Short bio
|
I was
born in Zgierz, Poland. In 2003, I received the M.Sc. degree in
Electronics and Telecommunication from the Technical University of Lodz
(PL), Poland. I was awarded two grants, one from the
Generalitat de Catalunya (FI) in 2004, and one from the Spanish
Ministry of
Education (FPU) in 2005-2008, for my Ph.D. work in the Mobile
Communication Research Group (GRCM) in the Department of Signal
Theory and Communication (TSC) at the Universitat Politecnica de
Catalunya (UPC) in Barcelona, Spain. During my Ph.D. studies I did two
short-term research visits. In 2006, I was a visiting scholar to the Department of Computer Science,
Karlstad University, Sweden, and in 2007, to the Protocol Engineering Labs (PEL) at
the University of Delaware, USA. I completed my Ph.D. in 2008. From
October 2008 I started working as a research fellow at the Hamilton
Institute, NUI Maynooth in Ireland. [Full biography] |
Research interests
|
|
Congestion control in TCP/IP networks is
traditionally handled using packet losses to indicate congestion [1].
This approach, called loss
based congestion control, proved to be working well over more
than 20 years. However, recent improvements in communication
technologies, and popularisation of wireless, high speed, and more
generally, links with high bandwidth delay product (BDP), has exposed
various shortcomings of loss-based congestion control. The main
criticism, apart from the increased loss rate is linked to
unefficient link utilisation and significant variations in network
delay.
An alternative approach to respond to congestion involves the use of network delay (delay based congestion control). This was first proposed by Jain already in 1989 [2] and since then, there has been much work and debate on this topic. Delay-based congestion control is conceptually very attractive. Potential benefits include the ability to allocate the network bandwidth between competing sources with: (i) low (zero) packet loss; (ii) very low queueing delay; and (iii) with full utilisation of network links. Networks which exhibit this property are said to operate at the knee of the throughput-delay curve [3]. Motivated by these and other potential benefits, delay-based congestion control remains an active area of research and new algorithms continue to be developed. Recent examples include: Fast TCP [4]; Microsoft Compound [5] (partially based on delay); and more recent delay-based additive increase multiplicative decrease (AIMD) variants [6-9]. Despite this large body of work, several issues concerning the use of queueing delay remain to be resolved before delay based congestion control can be deployed. These include: (i) the difficulty in obtaining delay estimates from network measurements [10]; (ii) network sampling issues [10–13]; (iii) the inability of existing delay-based algorithms to maintain a low standing queue [12, 13]; and (iv) the inability of delay-based flows to coexist fairly with loss-based flows in mixed environments. Our novel design Coexistant TCP (C-TCP) tries to tackle one of the biggest unresolved problems of the delay-based congestion control, namely, the coexistence problem. [More...] |
[1] V. Jacobson, “Congestion avoidance
and control,” in SIGCOMM’88:
Symposium proceedings on Communications architectures and protocols. New York, NY, USA: ACM, 1988, pp.
314–329.
[2] R. Jain, “A delay-based approach for congestion avoidance in interconnected heterogeneous computer networks,” ACM Comp. Commun. Rev., vol. 19, no. 5, pp. 56–71, 1989. [3] R. Jain and K. Ramakrishnan, “Congestion avoidance in computer networks with a connectionless network layer: concepts, goals and methodology,” Computer Networking Symposium, pp.134–143, 1988. [4] J. Wang, D. X. Wei, and S. Low, “Modelling and stability of FAST TCP,” in Proc. of the 24th IEEE INFOCOM Conference, vol. 2, 2005, pp. 938– 948. [5] K. Tan, J. Song, Q. Zhang, and M. Sridharan, “A compound TCP approach for high-speed and long distance networks,” 2005, Technical Report - Microsoft Research MSR-TR-2005-86. [6] D. Leith, J. Heffner, R. Shorten, and G. McCullagh, “Delay-based AIMD congestion control,” in Proc. of the 5th Int’l Workshop on Protocols for FAST Long-Distance Networks (PFLDnet), 2007. [7] D. Leith, R. Shorten, G. McCoullagh, L. Dunne, and F. Baker, “Making available base RTT for use in congestion control applications,” IEEE Commun. Lett., vol. 12, no. 6, pp. 429–431, June 2008. [8] S. Bhandarkar, A. Reddy, Y. Zhang, and D. Loguinov, “Emulating AQM from end hosts,” SIGCOMM Comput. Commun. Rev., vol. 37, no. 4, pp. 349–360, October 2007. [9] K. Kotla and A. Reddy, “Making a delay-based protocol adaptive to heterogeneous environments,” in Proc. of the 16th International Workshop on Quality of Service (IWQoS 2008), June 2008, pp. 100–109. [10] J. Martin, A. Nilsson, and I. Rhee, “Delay-based congestion avoidance for TCP,” IEEE/ACM Trans. Netw., vol. 11, no. 3, pp. 356–369, June 2003. [11] R. Prasad, M. Jain, and C. Dovrolis, “On the effectiveness of delay-based congestion avoidance,” in Proc. of the 2nd Int’l Workshop on Protocols for FAST Long-Distance Networks (PFLDnet), 2004. [12] G. McCullagh, “Exploring delay based TCP congestion control,” Master’s thesis, Hamilton Institute, NUI Maynooth, Ireland, 2008. [13] C. Kellett, R. Shorten, and D. Leith, “A review of delay-based congestion control,” 2006, Cisco Report - Distributed to project partners, March 2006. |
The Stream Control
Transmission Protocol (SCTP) was first announced in October 2000 in
the, now an obsolete, RFC 2960. Current protocol specification is
defined in the RFC 4960.
SCTP provides a reliable, full-duplex connection with flow and
congestion control algorithms that are derived from TCP, thus following
the same Additive Increase Multiplicative Decrease (AIMD) behavior. An
SCTP connection is called association, and is established using a
four-way handshake (instead of a three-way handshake as in TCP) in
order to improve protocol security and make it resistant to blind
denial of service (DoS) attacks (such as flooding and masquerade). What
made SCTP a subject of considerable interest however, are two new
features it introduces: multihoming
and multistreaming.
Multihoming (illustrated above) binds multiple source-destination IP addresses for a single association between two SCTP endpoints. These IP addresses are exchanged and verified during the association setup, and each destination transport address is considered as a different path towards the corresponding endpoint. Multihoming in SCTP was originally designed to increase the robustness of the signalling environments. Consequently, the scope of use for multihoming defined within the RFC 4960 is only for handling single retransmissions and performing primary path failover in case of a permanent link failure. Any other applications, e.g., transport-layer handover or loadsharing over multiple network paths, are not supported within the standard SCTP specification, and instead should be covered by dedicated protocol extensions (e.g., transport-layer handover application of multihoming requires Dynamic Address Reconfiguration (DAR) SCTP extension, defined in the RFC 5061). Regardless of this limitation, SCTP multihoming seems a promising protocol feature that may easily be leveraged to provide support for both mentioned applications. Multistreaming, the second of the new SCTP features (illustrated above), allows the establishment of associations with multiple streams. Streams are unidirectional data flows within a single association. The number of requested streams is declared during the association setup and the streams are valid during the entire association lifetime. Each stream is distinguished with the Stream Identifier field included in each chunk (SCTP packet consist of a common header and one or more chunks), so that chunks from different streams can be bundled inside one SCTP PDU. To preserve order within a stream the Stream Sequence Number (SSN) is used. Consequently, TCP’s HoL blocking problem stalling entire TCP connection is reduced to the affected SCTP stream only, as data received in order within a stream (handled by SSN) but not within the entire association (counted using TSN) can be delivered to the application. Among the most important applications of multistreaming are priority stream scheduling, preferential treatment, and reducing the latency of streaming multimedia in high-loss environments. An exhaustive collection of all SCTP-related research can be found here [SCTP survey]. |
Next generation mobile data
networks are expected to achieve a high degree of inter-networking so
that the mobile users can truly experience seamless access to their
services, irrespective of the radio technology being used. In such
scenarios, IP networking is becoming the keystone capable to turn this
vision into a reality. Hence,
mobility management solutions for IP networks are expected to provide
seamless mobility across multiple radio access options. Earlier works
on the mobility management problem discussed various solutions, mainly
in network and application layer of the ISO/OSI protocol stack. More
recently, transport layer handover schemes emerged, and are currently
receiving a notable attention in the research community, as they seem
to match very well the basic paradigm of the IP networking,
where intelligence is moved towards the edges of the network. My Ph.D. dissertation, and all mobility related articles, evaluate the concept of seamless handover management at the transport layer (in particular SCTP protocol) in IP-based heterogeneous wireless networks, as well as introduce a cross-layer design, and concurrent multipath transfer to optimize such a handover process. |
Publications | |||||
|
|||||
Congestion control | |||||
[1] | Delay based congestion
control for
heterogeneous environments, Ł. Budzisz, R. Stanojevic, A. Schlotte, R. Shorten and F. Baker submitted to IEEE/ACM Transactions on Networking magazine, under evaluation. For more results see [here] |
||||
[2] | On the fair
coexistence of loss- and delay-based TCP Ł. Budzisz, R. Stanojevic, A. Schlotte, R. Shorten and F. Baker Proc. of the 17th International Workshop on Quality of Service (IWQoS 2009), Charleston, SC, United States, July 2009. |
||||
[3] | A strategy for fair
coexistence of loss and delay-based
congestion control algorithms Ł. Budzisz, R. Stanojevic, R. Shorten and F. Baker IEEE Communications Letters, vol. 13, no. 7, p. 555-557, July 2009. |
||||
Mobility management | |||||
[1] | A taxonomy and survey
of SCTP research Ł. Budzisz, J. Garcia,
A. Brunstrom, and R. Ferrús
|
||||
[2] | On Concurrent Multipath
Transfer in SCTP-based handover scenarios Ł. Budzisz, R.
Ferrús, F. Casadevall, and P.Amer
|
||||
[3] | Design principles
and performance evaluation of mSCTP-CMT for transport-layer based
handover Ł. Budzisz, R.
Ferrús, and F. Casadevall
Proc. of the 69th IEEE Vehicular
Technology Conference (VTC 2009-Spring),
Barcelona, Spain, April 2009. |
||||
[4] |
Towards transport-layer
mobility: Evolution of SCTP multihoming Ł. Budzisz, R.
Ferrús, A. Brunstrom, R. Grinnemo, K.-J.
Fracchia, G. Galante, and F. Casadevall
|
||||
[5] |
An analytical
estimation of the failover time in SCTP
multihoming scenarios Ł. Budzisz, R.
Ferrús, K.-J. Grinnemo, A. Brunstrom, and F.
Casadevall
|
||||
[6] |
SCTP multihoming
performance in dynamically changing channels with the influence of
link-layer retransmissions Ł. Budzisz, R.
Ferrús, and F. Casadevall
Proc. of the 64th IEEE Vehicular
Technology Conference (VTC 2006-Fall),
pp. 2624–2628,
Montreal, Canada, September 2006. |
||||
[7] |
On the
performance of multihoming SCTP in dynamically changing radio
channels Ł. Budzisz, R.
Ferrús, and F. Casadevall
|
||||
[8] |
Study on
transport layer handover using SCTP Ł. Budzisz, R.
Ferrús, and F. Casadevall
|
||||
Network simulation and emulation tools | |||||
[1] |
NMLab: A Co-Simulation
Framework for Matlab and ns-2 O. Heimlich, R.
Sailer, and Ł. Budzisz
|
||||
[2] |
Evaluation of
Perceived QoS with Multimedia Applications in a Heterogeneous Wireless
Network N. Vucevic, F.
Bernardo, A. Umbert, and Ł. Budzisz
|
||||
[3] |
A beyond 3G
real-time testbed for an All-IP heterogeneous network F. Bernardo, N.
Vucevic, Ł. Budzisz, and A. Umbert
|
||||
[4] |
An all-IP
heterogeneous wireless testbed for RAT selection and e2e QoS
evaluation A. Umbert, Ł. Budzisz, F. Bernardo, and N. Vucevic
|
||||
[5] |
A Performance
Evaluation Framework of A Rate-Controlled MPEG
Video Transmission over UMTS Networks N. Akar, M. Barbera,
Ł. Budzisz, R.
Ferrús, E. Kankaya, and G. Schembra
|
version 1.1,
|