Transcript www.aui.ma
Computer Networks Set 10 X.25, ATM and Frame Relay X.25 Ñ 1976 Ñ Interface between host and packet switched network Ñ Almost universal on packet switched networks and packet switching in ISDN Ñ Defines three layers Ñ Physical Ñ Link Ñ Packet X.25 - Physical Ñ Interface between attached station and link to node Ñ Data terminal equipment DTE (user equipment) Ñ Data circuit terminating equipment DCE (node) Ñ Uses physical layer specification X.21 Ñ Reliable transfer across physical link Ñ Sequence of frames X.25 - Link, Packet Ñ Link Access Protocol Balanced (LAPB) Ñ Subset of HDLC Ñ Packet: External virtual circuits Ñ Logical connections (virtual circuits) between subscribers X.25 Use of Virtual Circuits Virtual Circuit Service Ñ Virtual Call Ñ Dynamically established Ñ Permanent virtual circuit Ñ Fixed network assigned virtual circuit Virtual Call Packet Format Multiplexing Ñ DTE can establish 4095 simultaneous virtual circuits with other DTEs over a single DTC-DCE link Ñ Packets contain 12 bit virtual circuit number Virtual Circuit Numbering Flow and Error Control Ñ HDLC (Chapter 7) Packet Sequences Ñ Complete packet sequences Ñ Allows longer blocks of data across network with smaller packet size without loss of block integrity Ñ A packets Ñ M bit 1, D bit 0 Ñ B packets Ñ The rest Ñ Zero or more A followed by B Reset and Restart Ñ Reset Ñ Reinitialize virtual circuit Ñ Sequence numbers set to zero Ñ Packets in transit lost Ñ Up to higher level protocol to recover lost packets Ñ Triggered by loss of packet, sequence number error, congestion, loss of network internal virtual circuit Ñ Restart Ñ Equivalent to a clear request on all virtual circuits Ñ E.g. temporary loss of network access ATM: Protocol Architecture Ñ Similarities between ATM and packet switching Ñ Transfer of data in discrete chunks Ñ Multiple logical connections over single physical interface Ñ In ATM flow on each logical connection is in fixed sized packets called cells Ñ Minimal error and flow control Ñ Reduced overhead Ñ Data rates (physical layer) 25.6Mbps to 622.08Mbps Protocol Architecture (diag) Reference Model Planes Ñ User plane Ñ Provides for user information transfer Ñ Control plane Ñ Call and connection control Ñ Management plane Ñ Plane management Ñ whole system functions Ñ Layer management Ñ Resources and parameters in protocol entities ATM Logical Connections Ñ Virtual channel connections (VCC) Ñ Analogous to virtual circuit in X.25 Ñ Basic unit of switching Ñ Between two end users Ñ Full duplex Ñ Fixed size cells Ñ Data, user-network exchange (control) and network-network exchange (network management and routing) Ñ Virtual path connection (VPC) Ñ Bundle of VCC with same end points ATM Connection Relationships Advantages of Virtual Paths Ñ Simplified network architecture Ñ Increased network performance and reliability Ñ Reduced processing Ñ Short connection setup time Ñ Enhanced network services Call Establishment Using VPs Virtual Channel Connection Uses Ñ Between end users Ñ End to end user data Ñ Control signals Ñ VPC provides overall capacity Ñ VCC organization done by users Ñ Between end user and network Ñ Control signaling Ñ Between network entities Ñ Network traffic management Ñ Routing VP/VC Characteristics Ñ Quality of service Ñ Switched and semi-permanent channel connections Ñ Call sequence integrity Ñ Traffic parameter negotiation and usage monitoring Ñ VPC only Ñ Virtual channel identifier restriction within VPC Control Signaling - VCC Ñ Done on separate connection Ñ Semi-permanent VCC Ñ Meta-signaling channel Ñ Used as permanent control signal channel Ñ User to network signaling virtual channel Ñ For control signaling Ñ Used to set up VCCs to carry user data Ñ User to user signaling virtual channel Ñ Within pre-established VPC Ñ Used by two end users without network intervention to establish and release user to user VCC Control Signaling - VPC Ñ Semi-permanent Ñ Customer controlled Ñ Network controlled ATM Cells Ñ Fixed size Ñ 5 octet header Ñ 48 octet information field Ñ Small cells reduce queuing delay for high priority cells Ñ Small cells can be switched more efficiently Ñ Easier to implement switching of small cells in hardware ATM Cell Format Header Format Ñ Generic flow control Ñ Only at user to network interface Ñ Controls flow only at this point Ñ Virtual path identifier Ñ Virtual channel identifier Ñ Payload type Ñ e.g. user info or network management Ñ Cell loss priority Ñ Header error control Generic Flow Control (GFC) Ñ Control traffic flow at user to network interface (UNI) to alleviate short term overload Ñ Two sets of procedures Ñ Uncontrolled transmission Ñ Controlled transmission Ñ Every connection either subject to flow control or not Ñ Subject to flow control Ñ May be one group (A) default Ñ May be two groups (A and B) Ñ Flow control is from subscriber to network Ñ Controlled by network side Single Group of Connections (1) Ñ Terminal equipment (TE) initializes two variables Ñ TRANSMIT flag to 1 Ñ GO_CNTR (credit counter) to 0 Ñ If TRANSMIT=1 cells on uncontrolled connection may be sent any time Ñ If TRANSMIT=0 no cells may be sent (on controlled or uncontrolled connections) Ñ If HALT received, TRANSMIT set to 0 and remains until NO_HALT Single Group of Connections (2) Ñ If TRANSMIT=1 and no cell to transmit on any uncontrolled connection: Ñ If GO_CNTR>0, TE may send cell on controlled connection Ñ Cell marked as being on controlled connection Ñ GO_CNTR decremented Ñ If GO_CNTR=0, TE may not send on controlled connection Ñ TE sets GO_CNTR to GO_VALUE upon receiving SET signal Ñ Null signal has no effect Use of HALT Ñ To limit effective data rate on ATM Ñ Should be cyclic Ñ To reduce data rate by half, HALT issued to be in effect 50% of time Ñ Done on regular pattern over lifetime of connection Two Queue Model Ñ Two counters Ñ GO_CNTR_A, GO_VALUE_A,GO_CNTR_B, GO_VALUE_B Header Error Control Ñ 8 bit error control field Ñ Calculated on remaining 32 bits of header Ñ Allows some error correction HEC Operation at Receiver Effect of Error in Cell Header Impact of Random Bit Errors Transmission of ATM Cells Ñ 622.08Mbps Ñ 155.52Mbps Ñ 51.84Mbps Ñ 25.6Mbps Ñ Cell Based physical layer Ñ SDH based physical layer Cell Based Physical Layer Ñ No framing imposed Ñ Continuous stream of 53 octet cells Ñ Cell delineation based on header error control field Cell Delineation State Diagram Impact of Random Bit Errors on Cell Delineation Performance Acquisition Time v Bit Error Rate SDH Based Physical Layer Ñ Imposes structure on ATM stream Ñ e.g. for 155.52Mbps Ñ Use STM-1 (STS-3) frame Ñ Can carry ATM and STM payloads Ñ Specific connections can be circuit switched using SDH channel Ñ SDH multiplexing techniques can combine several ATM streams STM-1 Payload for SDH-Based ATM Cell Transmission ATM Service Categories Ñ Real time Ñ Constant bit rate (CBR) Ñ Real time variable bit rate (rt-VBR) Ñ Non-real time Ñ Non-real time variable bit rate (nrt-VBR) Ñ Available bit rate (ABR) Ñ Unspecified bit rate (UBR) Real Time Services Ñ Amount of delay Ñ Variation of delay (jitter) CBR Ñ Fixed data rate continuously available Ñ Tight upper bound on delay Ñ Uncompressed audio and video Ñ Video conferencing Ñ Interactive audio Ñ A/V distribution and retrieval rt-VBR Ñ Time sensitive application Ñ Tightly constrained delay and delay variation Ñ rt-VBR applications transmit at a rate that varies with time Ñ e.g. compressed video Ñ Produces varying sized image frames Ñ Original (uncompressed) frame rate constant Ñ So compressed data rate varies Ñ Can statistically multiplex connections nrt-VBR Ñ May be able to characterize expected traffic flow Ñ Improve QoS in loss and delay Ñ End system specifies: Ñ Peak cell rate Ñ Sustainable or average rate Ñ Measure of how bursty traffic is Ñ e.g. Airline reservations, banking transactions UBR Ñ May be additional capacity over and above that used by CBR and VBR traffic Ñ Not all resources dedicated Ñ Bursty nature of VBR Ñ For application that can tolerate some cell loss or variable delays Ñ e.g. TCP based traffic Ñ Cells forwarded on FIFO basis Ñ Best efforts service ABR Ñ Application specifies peak cell rate (PCR) and minimum cell rate (MCR) Ñ Resources allocated to give at least MCR Ñ Spare capacity shared among all ARB sources Ñ e.g. LAN interconnection ATM Adaptation Layer Ñ Support for information transfer protocol not based on ATM Ñ PCM (voice) Ñ Assemble bits into cells Ñ Re-assemble into constant flow Ñ IP Ñ Map IP packets onto ATM cells Ñ Fragment IP packets Ñ Use LAPF over ATM to retain all IP infrastructure ATM Bit Rate Services Adaptation Layer Services Ñ Handle transmission errors Ñ Segmentation and re-assembly Ñ Handle lost and misinserted cells Ñ Flow control and timing Supported Application types Ñ Circuit emulation Ñ VBR voice and video Ñ General data service Ñ IP over ATM Ñ Multiprotocol encapsulation over ATM (MPOA) Ñ IPX, AppleTalk, DECNET) Ñ LAN emulation AAL Protocols Ñ Convergence sublayer (CS) Ñ Support for specific applications Ñ AAL user attaches at SAP Ñ Segmentation and re-assembly sublayer (SAR) Ñ Packages and unpacks info received from CS into cells Ñ Four types Ñ Type Ñ Type Ñ Type Ñ Type 1 2 3/4 5 AAL Protocols Segmentation and Reassembly PDU AAL Type 1 Ñ CBR source Ñ SAR packs and unpacks bits Ñ Block accompanied by sequence number AAL Type 2 Ñ VBR Ñ Analog applications AAL Type 3/4 Ñ Connectionless or connected Ñ Message mode or stream mode AAL Type 5 Ñ Streamlined transport for connection oriented higher layer protocols CPCS PDUs Example AAL 5 Transmission Frame Relay Ñ Designed to be more efficient than X.25 Ñ Developed before ATM Ñ Larger installed base than ATM Ñ ATM now of more interest on high speed networks ATM Traffic Ñ High speed, smallManagement cell size, limited overhead bits Ñ Still evolving Ñ Requirements Ñ Majority of traffic not amenable to flow control Ñ Feedback slow due to reduced transmission time compared with propagation delay Ñ Wide range of application demands Ñ Different traffic patterns Ñ Different network services Ñ High speed switching and transmission increases volatility Latency/Speed Effects Ñ ATM 150Mbps Ñ ~2.8x10-6 seconds to insert single cell Ñ Time to traverse network depends on propagation delay, switching delay Ñ Assume propagation at two-thirds speed of light Ñ If source and destination on opposite sides of USA, propagation time ~ 48x10-3 seconds Ñ Given implicit congestion control, by the time dropped cell notification has reached source, 7.2x106 bits have been transmitted Ñ So, this is not a good strategy for ATM Cell Delay Variation Ñ For ATM voice/video, data is a stream of cells Ñ Delay across network must be short Ñ Rate of delivery must be constant Ñ There will always be some variation in transit Ñ Delay cell delivery to application so that constant bit rate can be maintained to application Time Re-assembly of CBR Cells Network Contribution to Cell Delay Variation Ñ Packet switched networks Ñ Queuing delays Ñ Routing decision time Ñ Frame relay Ñ As above but to lesser extent Ñ ATM Ñ Less than frame relay Ñ ATM protocol designed to minimize processing overheads at switches Ñ ATM switches have very high throughput Ñ Only noticeable delay is from congestion Ñ Must not accept load that causes congestion Cell Delay Variation At The UNI Ñ Application produces data at fixed rate Ñ Processing at three layers of ATM causes delay Ñ Interleaving cells from different connections Ñ Operation and maintenance cell interleaving Ñ If using synchronous digital hierarchy frames, these are inserted at physical layer Ñ Can not predict these delays Origins of Cell Delay Variation Traffic and Congestion Control Framework Ñ ATM layer traffic and congestion control should support QoS classes for all foreseeable network services Ñ Should not rely on AAL protocols that are network specific, nor higher level application specific protocols Ñ Should minimize network and end to end system complexity Timings Considered Ñ Cell insertion time Ñ Round trip propagation time Ñ Connection duration Ñ Long term Ñ Determine whether a given new connection can be accommodated Ñ Agree performance parameters with subscriber Traffic Management and Congestion Control Techniques Ñ Resource management using virtual paths Ñ Connection admission control Ñ Usage parameter control Ñ Selective cell discard Ñ Traffic shaping Resource Management Using Virtual Paths Ñ Separate traffic flow according to service characteristics Ñ User to user application Ñ User to network application Ñ Network to network application Ñ Concern with: Ñ Cell loss ratio Ñ Cell transfer delay Ñ Cell delay variation Configuration of VCCs and VPCs Allocating VCCs within VPC Ñ All VCCs within VPC should experience similar network performance Ñ Options for allocation: Ñ Aggregate peak demand Ñ Statistical multiplexing Connection Admission Control Ñ First line of defence Ñ User specifies traffic characteristics for new connection (VCC or VPC) by selecting a QoS Ñ Network accepts connection only if it can meet the demand Ñ Traffic contract Ñ Peak cell rate Ñ Cell delay variation Ñ Sustainable cell rate Ñ Burst tolerance Usage Parameter Control Ñ Monitor connection to ensure traffic cinforms to contract Ñ Protection of network resources from overload by one connection Ñ Done on VCC and VPC Ñ Peak cell rate and cell delay variation Ñ Sustainable cell rate and burst tolerance Ñ Discard cells that do not conform to traffic contract Ñ Called traffic policing Traffic Shaping Ñ Smooth out traffic flow and reduce cell clumping Ñ Token bucket ATM-ABR Traffic Management Ñ Some applications (Web, file transfer) do not have well defined traffic characteristics Ñ Best efforts Ñ Allow these applications to share unused capacity Ñ If congestion builds, cells are dropped Ñ Closed loop control Ñ ABR connections share available capacity Ñ Share varies between minimum cell rate (MCR) and peak cell rate (PCR) Ñ ARB flow limited to available capacity by feedback Ñ Buffers absorb excess traffic during feedback delay Ñ Low cell loss Feedback Mechanisms Ñ Transmission rate characteristics: Ñ Allowed cell rate Ñ Minimum cell rate Ñ Peak cell rate Ñ Initial cell rate Ñ Start with ACR=ICR Ñ Adjust ACR based on feedback from network Ñ Resource management cells Ñ Congestion indication bit Ñ No increase bit Ñ Explicit cell rate field Variations in Allowed Cell Rate Cell Flow Rate Control Feedback Ñ EFCI (Explicit forward congestion indication) marking Ñ Relative rate marking Ñ Explicit rate marking Frame Relay Congestion Control Ñ Minimize discards Ñ Miantain agreed QoS Ñ Minimize probability of one end user monoply Ñ Simple to implement Ñ Little overhead on network or user Ñ Create minimal additional traffic Ñ Distribute resources fairly Ñ Limit spread of congestion Ñ Operate effectively regardless of traffic flow Ñ Minimum impact on other systems Ñ Minimize variance in QoS Techniques Ñ Discard strategy Ñ Congestion avoidance Ñ Explicit signaling Ñ Congestion recovery Ñ Implicit signaling mechanism Traffic Rate Management Ñ Must discard frames to cope with congestion Ñ Arbitrarily, no regard for source Ñ No reward for restraint so end systems transmit as fast as possible Ñ Committed information rate (CIR) Ñ Data in excess of this liable to discard Ñ Not guaranteed Ñ Aggregate CIR should not exceed physical data rate Ñ Committed burst size Ñ Excess burst size Operation of CIR Relationship Among Congestion Parameters Explicit Signaling Ñ Network alerts end systems of growing congestion Ñ Backward explicit congestion notification Ñ Forward explicit congestion notification Ñ Frame handler monitors its queues Ñ May notify some or all logical connections Ñ User response Ñ Reduce rate Frame Relay Background - X.25 Ñ Call control packets, in band signaling Ñ Multiplexing of virtual circuits at layer 3 Ñ Layer 2 and 3 include flow and error control Ñ Considerable overhead Ñ Not appropriate for modern digital systems with high reliability Frame Relay - Differences Ñ Call control carried in separate logical connection Ñ Multiplexing and switching at layer 2 Ñ Eliminates one layer of processing Ñ No hop by hop error or flow control Ñ End to end flow and error control (if used) are done by higher layer Ñ Single user data frame sent from source to destination and ACK (from higher layer) sent back Advantages and Disadvantages Ñ Lost link by link error and flow control Ñ Increased reliability makes this less of a problem Ñ Streamlined communications process Ñ Lower delay Ñ Higher throughput Ñ ITU-T recommend frame relay above 2Mbps Protocol Architecture Control Plane Ñ Between subscriber and network Ñ Separate logical channel used Ñ Similar to common channel signaling for circuit switching services Ñ Data link layer Ñ LAPD (Q.921) Ñ Reliable data link control Ñ Error and flow control Ñ Between user (TE) and network (NT) Ñ Used for exchange of Q.933 control signal messages User Plane Ñ End to end functionality Ñ Transfer of info between ends Ñ LAPF (Link Access Procedure for Frame Mode Bearer Services) Q.922 Ñ Frame delimiting, alignment and transparency Ñ Frame mux and demux using addressing field Ñ Ensure frame is integral number of octets (zero bit insertion/extraction) Ñ Ensure frame is neither too long nor short Ñ Detection of transmission errors Ñ Congestion control functions LAPF Core Formats User Data Transfer Ñ One frame type Ñ User data Ñ No control frame Ñ No inband signaling Ñ No sequence numbers Ñ No flow nor error control Required Reading Ñ Stallings Chapter 11 Ñ ATM Forum Web site Ñ Frame Relay forum