Constructions of Optical Priority Queues Jay Cheng Department of Electrical Engineering & Institute of Communications Engineering National Tsing Hua University.

Slides:



Advertisements
Similar presentations
Lecture 4. Topics covered in last lecture Multistage Switching (Clos Network) Architecture of Clos Network Routing in Clos Network Blocking Rearranging.
Advertisements

Optimal-Complexity Optical Router Hadas Kogan, Isaac Keslassy Technion (Israel)
Resource Allocation: Deterministic Analysis. Traffic Model Stochastic Different sample paths with different properties Expected case analysis Deterministic.
Lecture 12. Emulating the Output Queue So far we have shown that it is possible to obtain the same throughput with input queueing as with output queueing.
Lecture: 4 WDM Networks Design & Operation
TCP/IP Protocol Suite 1 Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Chapter 13 Introduction to the Transport.
1 Part III Markov Chains & Queueing Systems 10.Discrete-Time Markov Chains 11.Stationary Distributions & Limiting Probabilities 12.State Classification.
School of Information Technologies TCP Congestion Control NETS3303/3603 Week 9.
1 Estimating Shared Congestion Among Internet Paths Weidong Cui, Sridhar Machiraju Randy H. Katz, Ion Stoica Electrical Engineering and Computer Science.
All-Optical Header Recognition M. Dagenais Department of Electrical and Computer Engineering, University of Maryland, College Park, MD 20742, USA
Nick McKeown CS244 Lecture 6 Packet Switches. What you said The very premise of the paper was a bit of an eye- opener for me, for previously I had never.
What's inside a router? We have yet to consider the switching function of a router - the actual transfer of datagrams from a router's incoming links to.
Making Parallel Packet Switches Practical Sundar Iyer, Nick McKeown Departments of Electrical Engineering & Computer Science,
ECS 152A Acknowledgement: slides from S. Kalyanaraman & B.Sikdar
Crossbar Switches Crossbar switches are an important general architecture for fast switches. 2 x 2 Crossbar Switches A general N x N crossbar switch.
April 10, HOL Blocking analysis based on: Broadband Integrated Networks by Mischa Schwartz.
Channel Allocation for the GPRS Design and Performance Study Huei-Wen Ferng, Ph.D. Assistant Professor Department of Computer Science and Information Engineering.
Analysis of a Packet Switch with Memories Running Slower than the Line Rate Sundar Iyer, Amr Awadallah, Nick McKeown Departments.
Katz, Stoica F04 EECS 122: Introduction to Computer Networks Performance Modeling Computer Science Division Department of Electrical Engineering and Computer.
Scaling Internet Routers Using Optics Producing a 100TB/s Router Ashley Green and Brad Rosen February 16, 2004.
Guaranteed Smooth Scheduling in Packet Switches Isaac Keslassy (Stanford University), Murali Kodialam, T.V. Lakshman, Dimitri Stiliadis (Bell-Labs)
October 5th, 2005 Jitter Regulation for Multiple Streams David Hay and Gabriel Scalosub Technion, Israel.
Chain: Operator Scheduling for Memory Minimization in Data Stream Systems Authors: Brian Babcock, Shivnath Babu, Mayur Datar, and Rajeev Motwani (Dept.
Little’s Theorem Examples Courtesy of: Dr. Abdul Waheed (previous instructor at COE)
1 The Designs and Analysis of a Scalable Optical Packet Switching Architecture Speaker: Chia-Wei Tuan Adviser: Prof. Ho-Ting Wu 3/4/2009.
Fundamental Complexity of Optical Systems Hadas Kogan, Isaac Keslassy Technion (Israel)
Maximum Size Matchings & Input Queued Switches Sundar Iyer, Nick McKeown High Performance Networking Group, Stanford University,
1 Achieving 100% throughput Where we are in the course… 1. Switch model 2. Uniform traffic  Technique: Uniform schedule (easy) 3. Non-uniform traffic,
“On the Integration of MPEG-4 streams Pulled Out of High Performance Mobile Devices and Data Traffic over a Wireless Network” Spyros Psychis, Polychronis.
Queuing Networks: Burke’s Theorem, Kleinrock’s Approximation, and Jackson’s Theorem Wade Trappe.
Power saving technique for multi-hop ad hoc wireless networks.
Analysis of Input Queueing More complex system to analyze than output queueing case. In order to analyze it, we make a simplifying assumption of "heavy.
7/15/2015HY220: Ιάκωβος Μαυροειδής1 HY220 Schedulers.
CS144, Stanford University Error in Q3-7. CS144, Stanford University Using longest prefix matching, the IP address will match which entry? a /8.
Load Balancing and Switch Scheduling Xiangheng Liu EE 384Y Final Presentation June 4, 2003.
Localized Asynchronous Packet Scheduling for Buffered Crossbar Switches Deng Pan and Yuanyuan Yang State University of New York Stony Brook.
Buffer Management for Shared- Memory ATM Switches Written By: Mutlu Apraci John A.Copelan Georgia Institute of Technology Presented By: Yan Huang.
Lesson 11: Solved M/G/1 Exercises
Merits of a Load-Balanced AAPN 1.Packets within a flow are transported to their correct destinations in sequence. This is due to the 1:1 logical connection.
“A non parametric estimate of performance in queueing models with long-range correlation, with applications to telecommunication” Pier Luigi Conti, Università.
Network Analysis A brief introduction on queues, delays, and tokens Lin Gu, Computer Networking: A Top Down Approach 6 th edition. Jim Kurose.
ATM SWITCHING. SWITCHING A Switch is a network element that transfer packet from Input port to output port. A Switch is a network element that transfer.
1 Copyright © Monash University ATM Switch Design Philip Branch Centre for Telecommunications and Information Engineering (CTIE) Monash University
Time Parallel Simulations II ATM Multiplexers and G/G/1 Queues.
Survey of Performance Analysis on Banyan Networks Written By Nathan D. Truhan Kent State University.
Analysis of a Protocol for Dynamic Configuration of IPv4 Link Local Addresses Using Uppaal Miaomiao Zhang Frits W. Vaandrager Department of Computer Science.
Univ. of TehranAdv. topics in Computer Network1 Advanced topics in Computer Networks University of Tehran Dept. of EE and Computer Engineering By: Dr.
Logical Topology Design
The Physical Layer Lowest layer in Network Hierarchy. Physical transmission of data. –Various flavors Copper wire, fiber optic, etc... –Physical limits.
Competitive Queue Policies for Differentiated Services Seminar in Packet Networks1 Competitive Queue Policies for Differentiated Services William.
1 Optical Packet Switching Techniques Walter Picco MS Thesis Defense December 2001 Fabio Neri, Marco Ajmone Marsan Telecommunication Networks Group
ISLIP Switch Scheduler Ali Mohammad Zareh Bidoki April 2002.
Nick McKeown Spring 2012 Lecture 2,3 Output Queueing EE384x Packet Switch Architectures.
Guaranteed Smooth Scheduling in Packet Switches Isaac Keslassy (Stanford University), Murali Kodialam, T.V. Lakshman, Dimitri Stiliadis (Bell-Labs)
Loss-Bounded Analysis for Differentiated Services. By Alexander Kesselman and Yishay Mansour Presented By Sharon Lubasz
1 Algorithms Queues, Stacks and Records stored in Linked Lists or Arrays.
Buffered Crossbars With Performance Guarantees Shang-Tse (Da) Chuang Cisco Systems EE384Y Thursday, April 27, 2006.
Non-Preemptive Buffer Management for Latency Sensitive Packets Moran Feldman Technion Seffi Naor Technion.
Courtesy Piggybacking: Supporting Differentiated Services in Multihop Mobile Ad Hoc Networks Wei LiuXiang Chen Yuguang Fang WING Dept. of ECE University.
Scheduling for QoS Management. Engineering Internet QoS2 Outline  What is Queue Management and Scheduling?  Goals of scheduling  Fairness (Conservation.
Univ. of TehranIntroduction to Computer Network1 An Introduction to Computer Networks University of Tehran Dept. of EE and Computer Engineering By: Dr.
Data Structures Using C, 2e
Queues.
Intro to Deterministic Analysis
Intro to Deterministic Analysis
Empirically Characterizing the Buffer Behaviour of Real Devices
Intro to Deterministic Analysis
Brian Babcock, Shivnath Babu, Mayur Datar, and Rajeev Motwani
Buffer Management for Shared-Memory ATM Switches
Chapter 2 Switching.
Presentation transcript:

Constructions of Optical Priority Queues Jay Cheng Department of Electrical Engineering & Institute of Communications Engineering National Tsing Hua University Hsinchu, Taiwan 30013, R.O.C. The slides are made by using MS PowerPoint 2007

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

Motivation As photons cannot be easily stopped, stored, and forwarded, the only known way to store optical packets without converting them into other media is to direct optical packets through optical Switches and fiber Delay Lines (SDL) so that packets appear at the right place and at the right time Recent advances in SDL constructions have shown that there exist systematic methods for constructing various types of optical queues, including buffered switches, FIFO multiplexers, FIFO queues, LIFO queues, priority queues, linear compressors, non-overtaking delay lines, and flexible delay lines The purpose of the slides is to help the audience understand our SDL constructions of priority queues (for simplicity, only the single-input single-output case is presented here). One of the well-recognized key problems for all-optical packet-switched networks is the lack of optical buffer for contention resolution among packets competing for the same resources The key idea in the SDL constructions is to design different types of optical buffers for different types of network elements that do not need buffers with random-access capability

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

Basic Assumptions Time in all optical links is slotted and synchronized so that a packet can be transmitted within a time slot. Packets are of the same size. A fiber delay line with delay d is an optical link that requires d time slots for a packet to traverse through. An M x M crossbar switch is a network element with M input links and M output links that realizes all the M! permutations between its inputs and outputs.

Control input link Arrival link Buffer size Departure link Loss link Definition of Priority Queues Priority queue with buffer B B a(t) c(t) d(t) (t) (P1) Flow conservation (P2) Non-idling (P3) Priority departure (P4) Maximum buffer usage (P5) Priority loss Arriving packets from the arrival link are either stored in the buffer or transmitted through the departure link or the loss link. If c(t)=1, then there is always a departing packet if there are packets in the buffer or there is an arriving packet. If there is a departing packet at time t, then the departing packet is the one with the highest priority among all the packets in the queue at time t. If c(t)=0, then there is a lost packet only when buffer is full and there is an arriving packet. If there is a lost packet at time t, then the lost packet is the one with the lowest priority among all the packets in the queue at time t.

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

The Operations of a Priority Queue (B=9) c(t) RAM

The Operations of a Priority Queue (B=9) t = 1 c(t)c(t)=0

The Operations of a Priority Queue (B=9) (P1) Flow conservation: Arriving packets from the arrival link are either stored in the buffer or transmitted through the departure link or the loss link c(t)=0 1 t = 1

The Operations of a Priority Queue (B=9) t = 1t = (P2) Non-idling: If c(t)=1, then there is always a departing packet if there are packets in the buffer or there is an arriving packet (P3) Priority departure: If there is a departing packet at time t, then the departing packet is the one with the highest priority among all of the packets in the queue at time t c(t)=0c(t)=1

The Operations of a Priority Queue (B=9) t = 2t = c(t)=1 c(t)=0

The Operations of a Priority Queue (B=9) t = 3t = c(t)=0

The Operations of a Priority Queue (B=9) t = 4t = c(t)=0

The Operations of a Priority Queue (B=9) t = 5t = c(t)=0

The Operations of a Priority Queue (B=9) t = 6t = c(t)=0

The Operations of a Priority Queue (B=9) t = 7t = c(t)=0

The Operations of a Priority Queue (B=9) t = 8t = c(t)=0

The Operations of a Priority Queue (B=9) t = 9t = c(t)=0

The Operations of a Priority Queue (B=9) t = 10t = (P4) Maximum buffer usage: If c(t)=0, then there is a lost packet only when buffer is full and there is an arriving packet (P5) Priority loss: If there is a lost packet at time t, then the lost packet is the one with the lowest priority among all of the packets in the queue at time t 7 8 c(t)=0

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

Our Previous SDL Construction Theorem If for all, then the following construction can be operated as a priority queue with buffer

An Example of Our Previous SDL Construction (B=9) t = 0 t = 1 (P1) Flow conservation is satisfied c(t) = 0 The key is to view empty time slots as fictitious packets that have priorities lower than those of real packets. Moreover, the priorities among the fictitious packets are decreasing in the order of their arrival times. Note that both the packet with the highest priority and the packet with the lowest priority appear at the inputs of the sorter at the beginning of each time slot.

An Example of Our Previous SDL Construction (B=9) t = 1t = 2 (P2) Non-idling and (P3) Priority departure are satisfied c(t) = 0c(t) =

An Example of Our Previous SDL Construction (B=9) t = 2t = 3 c(t) =1 c(t) =

An Example of Our Previous SDL Construction (B=9) t = 3 t = 4 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 4t = 5 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 5t = 6 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 6t = 7 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 7t = 8 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 8t = 9 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 9t = 10 c(t) = 0

An Example of Our Previous SDL Construction (B=9) t = 10t = 11 (P4) Maximum buffer usage and (P5) Priority loss are satisfied c(t) = 0

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

(A2) When a packet enters the i th delay line, there are at least d i -1 packets with priorities higher than its priority, and there are at least d i -1 packets with priorities lower than its priority An Improved SDL Construction The key to the success of our previous SDL construction is we make sure that the following condition is satisfied: (A1) Both the highest priority packet and the lowest priority packet appear at the inputs of the sorter at the beginning of each time slot Our improvement is made possible by the observation that the following condition (A2) implies that (A1) holds: In other words, all the packets stored in the fiber delay lines cannot be either the highest priority packet or the lowest priority packet until they appear at the inputs of the sorter

An Improved SDL Construction

The Concept of the Cells of the Fiber Delay Lines m m M-2m d 1 =1 d 2 =2 d m =m d M+1-m =m d M =1 d M-1 =2

The (i,j) th cell, i=1,2,…M, j=1,2,…,d i, is the j th cell of the i th fiber delay line. Example (M=7, m=3) (4,4) (1,1) (2,2)(2,1) (3,3)(3,2)(3,1) (4,3)(4,2)(4,1) (7,1) (6,2)(6,1) (5,3)(5,2)(5,1) i=1 i=2 i=3 i=4 i=5 i=6 i=7 m=3 M-2m=1 j=1 j=3 j=4 j=2 We view a fiber delay line with delay d as a sequential buffer that consists of d cells with each cell capable of holding a packet. A packet entering the i th delay line at time t will be stored in the (i,j) th cell at time t+j-1.

Tagged packet We consider a packet that enters the i th delay line at some time t 0 and call it the tagged packet 0 Let p i,j (t)= 1, if the priority of the packet in cell (i,j) at time t is higher than or equal to the priority of the tagged packet 0, otherwise. Properties of p i,j (t)

At time t-1 Properties of p i,j (t) At time t-1 At time t i= Monotonically decreasing

The total number of packets in the cells of the j th column that have properties higher than or equal to that of the tagged packet at time t

At time t-1 At time t

The total number of packets in the cells of the j th column that have properties higher than or equal to that of the tagged packet at time t

At time t-1 At time t

At time t-1 At time t p 1 (t) is the total number of packets in the first column at time t that have priorities higher than or equal to that of the tagged packet These M packets can only come from the arriving packet at time t and those packets stored at time t-1 in the cells (i,i) and (M+1-i,i) for i=1,2,…,m, and (i,d i ) for i=m+1,m+2,…,M-m Arriving packet M-2m

Outline Motivation Definition of Priority Queues The Operations of a Priority Queue (Electronic Random Access Memory (RAM) is used to Buffer Packets) Our Previous SDL Construction An Improved SDL Construction Conclusion and Future Works

We have considered the construction of an optical priority queue with a single (M+1) × (M+1) switch and M fiber delay lines, and have shown that such a construction can be operated as a priority queue with O(M 3 ) buffer size Further studies are needed to go beyond the O(M 3 ) buffer size and march toward the exponential upper bound O(2 M ) for a construction of an optical priority queue with a single (M+1) × (M+1) switch and M fiber delay lines

Thank you!