data structures for bandwidth reservations and quality of service on the internet andreas nilsson...

26
Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department of Computer Science and Electrical Engineering Luleå University of Technology

Upload: juniper-henry

Post on 29-Dec-2015

220 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Data structures for Bandwidth Reservations and Quality of

Service on the Internet

Andreas Nilsson

Division of Computer Science and Networking

Department of Computer Science and Electrical Engineering

Luleå University of Technology

Page 2: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Presentation outline• Paper presentation.• Background.• Queuing.• Bandwidth reservation problem (Two solutions):

– Static.

– Dynamic.

• Further Work

Page 3: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Introduction - papers

1. “Extended Expedited Forwarding: In-Time PHB group”.

2. “A Static Data Structure for Discrete Advance Bandwidth Reservations on the Internet”.

3. “Performance of QoS Agents for Provisioning Network Resources”.

4. “Data Structure for a Time-Based Bandwidth Reservations Problem”.

5. “Application of resource allocation problem to the analysis of spectra”.

Page 4: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Internet• Largest network of networks.

– 1977-111, 1987-10000, 2002 hosts.

• Packet switched network.– All packets are addressed.

• No guarantees.– Packets are delivered at best effort.

T O : C lien t 2F R O M : C lien t 1

8102

Page 5: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Real time on the Internet

• For video, TV, phone, process control etc.• Demands packets:

– At constant delay.

– In order.

– At a specified bandwidth.

• A “circuit switched Internet” is the “goal”.

Page 6: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Queuing

• Conforming traffic must be delivered In-Time and in order therefore a dedicated “lane” through the routers is needed.

• Excess traffic (additional traffic improving the

quality) belongs to a conforming flow, and must be delivered in order and in time, but may be tossed due to congestion.

• Best effort, the background traffic.• Space complexity: O(n) Time: O(1) per

packet

C o n fo rm in g

E x cess

B est E ffo rt

Page 7: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Reservations and Quality of Service

• Provide different levels of service on the Internet for different applications.– Thorax surgical robot controlled over the Internet.

– Live TV over the Internet.

– File transfer (web) over the Internet.

• Olov Schelén et.al. presented a QoS providing architecture, called Bandwidth Brokers.– Clients make bandwidth reservations for the traffic

through the Internet.

Page 8: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Bandwidth brokers• They manage the bandwidth reservations on

one link.

N etw o rk 1 N e tw o rk 2

B an d w id th b ro k e r

Page 9: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Interaction between bandwidth brokers

• Usually the route between two computers consists of several links.

• Entire route has to be reserved.

B an d w id th b ro k e r

N e tw o rk 4

B an d w id th b ro k e r

N e tw o rk 3

B an d w id th b ro k e r

N e tw o rk 1 N e tw o rk 2

B an d w id th b ro k e r

N e tw o rk 5

Page 10: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Bandwidth Reservation Problem• The problem of how to make the data structure needed

to handle the reservations in the Bandwidth Brokers efficient.

• Presents two solutions, in the comparison based model of computation:– Static (AST).– Dynamic (BinSeT).

• Both solutions implement the same set of operations:– Insert(Bandwidth, Interval)– Delete(Bandwidth, Interval)– MaxReserved(Interval)Bandwidth

Page 11: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Bandwidth Reservation ProblemAn example

Page 12: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Advanced Segment Tree (AST)

• Modified segment tree.• The tree is static, i.e.:

– The tree is only built once, and in advance, no nodes are added or deleted and therefore the tree is always perfectly balanced and covers a bounded universe M.

– The intervals available for reservations are determined upon construction

• All nodes on a level l:– have the same number of children.– represent a time interval.– cover intervals of the same size.– have intervals that mutually don't intersect.– have intervals, the union of which is M.– have an interval that is contained within one interval on level l’, where l’<l.

Page 13: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Tim e

n o d e v a lu e = 0m ax v a lu e = = 2 2 0m a x(0 + 3 0 ,2 0 + 2 0 0 )

n o d e v a lu e = 6 0m ax v a lu e = 9 0

9 00

00

00

00

00

3 00

2 0 00

n o d e v a lu e = 0m ax v a lu e = 3 0

n o d e v a lu e = 2 0m ax v a lu e = 2 0 0

n o d e v a lu e = 1 0m ax v a lu e = = 2 5 0m a x(5 0 + 2 0 0 ,0 + 2 2 0 )

n o d e v a lu e = 5 0m ax v a lu e = = 2 0 0m a x(6 0 + 9 0 ,1 2 0 + 8 0 )

n o d e v a lu e = 1 2 0m ax v a lu e = 8 0

8 00

AST - Nodes and their values

Page 14: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Binary Segment Tree (BinSeT)• Data structure that combines the properties of a binary search tree

and a segment tree.• The tree is dynamic, i.e.:

– The tree is continuously constructed, nodes are added upon insertions. Therefore the tree has to be balanced continuously.

– The intervals are chosen completely freely.– The tree covers an unbounded universe

• Consists of nodes (internal nodes) and leaves (nodes)– Nodes store

, max reserved bandwidth below the node , change during the interval , split time of the node

– Leaves store , max reserved bandwidth below the leaf , change at time t , start time that the leaf represents

Page 15: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

A simple example BinSeT

1

1

B a n d w id th

T im e2

2

3 4-1

-2

0

1

1

B a n d w id th

T im e2

2

3 4-1

-2

0

1

1

B a n d w id th

T im e2

2

3 4-1

-2

0

Tim e

B an d w id th

3

4

1

2

1 2 3 40

Tim e

B an d w id th

3

4

1

2

1 2 3 40

Tim e

B an d w id th

3

4

1

2

1 2 3 40

23

+ 20

0

3

+ 10

-10

-20

-3

0

22

+ 2 -1 -100

0

1

0

-10

01

0+ 1

Page 16: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Further work

• Policing for the Bandwidth Brokers.• Generalisation of the data structures to more

dimensions.• A modified BinSeT to efficiently find unreserved

resources over time.• Other queries based on operations defining

monoids.• Time based data structures (similar way of reasoning as in

A. Brodnik, S. Carlsson, J. Karlsson, and J. I. Munro, “Worst Case Constant Time Priority Queue”, SODA 2001)

Page 17: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

AST versus BinSeT

Data-structure

Comment Space Time

query

Time

updateAST static, wrap,

implicitO(M) O(log M) O(log M)

BinSeT Dynamic O(n) O(log n) O(log n)

M = size of universen = number of unique interval end points (event points)

Page 18: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Time for questions

[email protected]

http://www.sm.luth.se/~andreas/publications/

Thank you

Page 19: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

max(1,1 )=3+2

max(1,1+1)=2

max( 1, )= 1 1 1

max( , -6)=4 4 4

max( ,1 2)=1 + 3

max(1,1-3)=1

max(-2,-2+1)=-1

max(1,1-2)=1

max(7,1 )=7+4

max(5,5 )=7+2

max(4,1+0)=4

max(3,3 )=5+2

max( 1, )=2 2 4+

max(3,3+1)=4

max(-1,-1+1)=0

+1

+2

+1

+1 -1

-1

+4

-6

+1

+2

+1

-3

-2

+1

+1 -2

Example 1 BinSeT

Page 20: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

max(7,1 )=7+4

max(4,1+0)=4

max(3,3+1)=4

max(-1,-1+1)=0

AC

C

A

A

A

B

C

C

C

A

....

Example 2 BinSeT

Page 21: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Paper 1 A Static Data Structure for Discrete Advance

Bandwidth Reservations on the Internet

• Publications:– Technical report 2003.

– Conference records of SNCNW 2003.

• Authors:– Andrej Brodnik and Andreas Nilsson.

• Description:– Shows a static data structure for bandwidth

reservations.

Page 22: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Paper 2 Performance of QoS Agents for Provisioning Network Resources

• Publications:– Conference records of IWQoS’99.

• Authors:– Olov Schelén, Andreas Nilsson, Joakim Norrgård and

Stephen Pink.

• Description:– Shows that the hardware requirements of a QoS agent

are quite modest by in simulations compare two algorithms for making reservations.

Page 23: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Paper 3Extended Expedited Forwarding:

In-Time PHB group

• Publications:– Technical Report 2003.– Conference records of ISCC’2003, FTC’2001 and

CSEE’2001.

• Authors:– Ulf Bodin, Andrej Brodnik, Johan Karlsson, Andreas

Nilsson and Olov Schelén.

• Description:– Shows a proposal on how to handle excess traffic in

queues in relation to conforming traffic and best effort traffic.

Page 24: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Paper 4Data Structure for a Time-Based Bandwidth Reservations Problem

• Publications:– Technical Report 2003.

– Conference records of CSEE’2002.

• Authors:– Andrej Brodnik and Andreas Nilsson.

• Description:– Shows a dynamic solution to the bandwidth reservation

problem.

Page 25: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Paper 5Application of resource allocation problem to the analysis of spectra

• Publications:– In proceedings of MATH/CHEM/COMP/2002.

• Authors:– Andrej Brodnik, Ante Graovac, Andreas Nilsson,

Tomaz Pisanski.

• Description:– Only an abstract for a presentation in which it is shown

that the solution to the Bandwidth reservation problem also is applicable when analyzing spectra.

Page 26: Data structures for Bandwidth Reservations and Quality of Service on the Internet Andreas Nilsson Division of Computer Science and Networking Department

Roberto Grossi

• Personal Data– Associate Professor at the University of Pisa.

– Ph.D. and Laurea in Computer Science at the University of Pisa.

– Member of the Algorithms and Data Structures group at the Department of Computer Science.

– Formerly, Assistant Professor at the University of Florence.