Chapter 2: High Speed Networks PDF

Chapter 2: High Speed Networks © 2024 Burkhard Stiller M02-1  Content  Topics – Communication architectures – Deficiencies and gaps – Services, protocols, phases – Service ca...

Chapter 2: High Speed Networks © 2024 Burkhard Stiller M02-1  Content  Topics – Communication architectures – Deficiencies and gaps – Services, protocols, phases – Service categories – Architectural view – Transfer modus – Switching: terminology, circuit, packet – Characteristics of High Speed Networks (HSN)  Objectives – To describe the major basis of architectures and transmission approaches – To discuss selected examples and pick the high-speed reasoning – To explain how underlying network transmission mechanisms serve applications © 2024 Burkhard Stiller M02-2  Communication Architectures  Applications require services  Networks provide service  Question: What does interconnect sufficiently Application services required and provided? Communication Subsystem (CS) – A communication subsystem Network  Problems: (Ideal situation) – Are network services useful for applications? – Is the communication subsystem efficient? – What are the influences? © 2024 Burkhard Stiller M02-3  ISO/OSI Basic Reference Model – Concept of hierarchical layering – Protocol Data Units (PDU) exchanged between peers – Service Data Units (SDU) exchanged between layers – Interface Data Units (IDU) exchanged between interfaces Application Application specific issues 7 Presentation Data formats 6 Session Dialogue control 5 Transport End-to-end Exchange 4 Network Routing on end-system basis 3 Link Link control 2 Physical Transmission 1 © 2024 Burkhard Stiller M02-4  Tasks of BRM Layers (1)  Application Layer – Handle the interface between the communication (sub-)system and applications – Functional standards, e.g., transaction processing (TP), file access and management (FTAM), or virtual terminal (VT)  Presentation Layer – Establish a common syntax for commands – Provide an agreed upon structure for data exchange – Encryption and decryption  Session Layer – Establish a common context for dialog control – Identify the present message synchronization © 2024 Burkhard Stiller M02-5  Tasks of BRM Layers (2)  Transport Layer – Establish and terminate end-to-end connections – Flow control and error control – Multiplexing  Network Layer – Interconnect different networks (routing and addressing) – Congestion control  Data Link Layer – Establish and terminate hop-by-hop connections – Flow control and bit error control (checksum)  Physical Layer – Define physical bit transmission media and devices © 2024 Burkhard Stiller M02-6  Deficiencies of the BRM  Communication between layers (inter-lay-com) – SDUs, layer-to-layer flow-control, and multiplexing  Communication between instances (intra-lay-com) – Set-up of connections, 2 * N control packets  Transparency – Redundant functionality, e.g., checksum or flow-control  Complex sub-layering – Additional layers, such as MAC/LLC on layer 2, and – Separation of user data transfer, signaling, management.  Limited selection of Quality-of-Service  Missing service integration – Connectionless or connection-oriented interfaces © 2024 Burkhard Stiller M02-7  The Service Gap  The “narrow” Communication Subsystem is due to two reasons: – Performance – Functionality Audio, Video, Data ISO/OSI  High-speed applications Layers require high-speed nets 7 Application  Low delay applications require low delay services 3–6 CS  Isochronous applications 2b require appropriate jitter 1 – 2a Network  Loss-intolerant applications (Real situation) require well suited error-control CS: Communication Subsystem © 2024 Burkhard Stiller M02-8  3-Component Model  Topic-dependent structure and coarse-grained layering  High flexibility in terms of additional services, less interfaces, and finer functional granularity within components  Separated protocol functions per each of the 3 components – Applications Entities within an end-system allow for handling a certain task that delivers specified services to users – Communication subsystem Application- Part of an end-system that processes communication- oriented relevant tasks, e.g., communication protocol processing Communication- Based on existing end-systems (hardware: computer) relevant including operating systems (software) – Networks Network- The infrastructure to transport information in form of dependent packets, cells, or byte or bit streams Examples as mentioned above © 2024 Burkhard Stiller M02-9  Service and Protocol  A service is defined at a Service Access Point via Service Primitives to be accessed by a Service User  A protocol is specified for a Service Provider using Service Primitives to stimulate control mechanisms (Caller) Service User Service User (Callee) Service Service Primitives Service Access Points Primitives Service Provider © 2024 Burkhard Stiller M02-10  Communication Services  General definition – “A service is defined by features and capabilities between two general objects that are related to each other.”  ISO definition (ISO/OSI Basic Reference Model) – “An (N)-service is the capability of the (N+1)-layer and the layer beneath it, which is provided to (N+1)-entities at the boundary between the (N)-layer and the (N+1)-layer.”  Service types – Connectionless or connection-oriented – Acknowledged or non-acknowledged  Service primitives – Request, indication, response, and confirmation © 2024 Burkhard Stiller M02-11  Service Phases  Connectionless service – Data transfer phase  Connection-oriented service – Connection set-up phase Synonyms: establish, create, initiate, connect – Data transfer phase – Connection tear-down phase Synonyms: terminate, close, release, disconnect  In case of an “abstract” connection – not being related to a specific layer – the wording of an “association” (in contrast to the ISO/OSI BRM terminology) is used © 2024 Burkhard Stiller M02-12  Communication Protocols  General definition – “A protocol is a set of rules to exchange information between entities in order to provide a service.”  ISO definition (ISO/OSI Basic Reference Model) – “A set of rules and formats (semantic and syntactic) which determine the communication behavior of entities in the performance of functions.”  Multimedia and high-performance definition – “A protocol is a set of rules to exchange information with high-performance between distributed entities across high-performance communication subsystems and networks in order to provide an efficient and specified multimedia service.” © 2024 Burkhard Stiller M02-13  Service Categories  Four main categories (transmission link) – Synchronous – Asynchronous – Isochronous – Plesiochronous  Two main categories (application view)  later chapters – Best effort service – Guaranteed service Statistic Deterministic  Independent structuring of these categories © 2024 Burkhard Stiller M02-14  Synchronous Service  User data are transmitted in data elements, such as – Cells, Packets – Messages – Protocol data units  Consecutive data units follow a strict timing. – Exact inter-element times – No variance of this time at all  A maximum end-to-end delay is defined inter-element time time t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 © 2024 Burkhard Stiller M02-15  Asynchronous Service  Consecutive data units do not follow a strict timing – Inter-element times may vary – Variance of this time does not follow any rules time t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 © 2024 Burkhard Stiller M02-16  Jitter (1)  Jitter or delay jitter is defined as the variance of the inter- element time between consecutive elements  Jitter is introduced by networks in data flows due to variable transmission delays Sender time t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 Network Jitter (+) Jitter (–) Receiver time t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 © 2024 Burkhard Stiller M02-17  Jitter (2)  Reasons for jitter (variable network delays) – Variable service times of intermediate nodes, such as routers or switches – Use of different data paths for data elements of a single flow  Even sending data continuously does not avoid jitter  Solution – If a maximum end-to-end delay is guaranteed and a maximum buffer length at the receiver is available, the jitter can be compensated. © 2024 Burkhard Stiller M02-18  Isochronous Service  Consecutive data units do not follow a strict, but a bounded timing – Inter-element time may vary according to maximum/minimum (absolute) values – Variance of this time does follow defined rules Maximum √ √ √ Jitter exceeded ! maximum Jitter √ time t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10  Definition of a maximum jitter derives a maximum and minimum delay for all elements, if sender sends continuously at a fixed rate  Isochronous services may be offered at – Service interface of the network-dependent component – Service interface of the communication-relevant component, e.g., a transport service  Isochronous services at transport may be provided even w/o isochronous transmissions M02-19 © 2024 Burkhard Stiller  Architectural View  Lower layers of an architecture comprise the physical layer and the transmission media in addition to the access to a network  Local Area Networks (LAN) subdivide the data link layer due to a variety of different access methods, such as token ring, carrier sense, or FDDI  Modern Wide Area Networks (WAN) take a different approach ISO/OSI BRM LAN WAN Logical Link Control Data Link Layer Transfer Modi Medium Access Physical Layer Physical Physical © 2024 Burkhard Stiller M02-20  Transfer Modi  A transfer modus describes a network technique to handle the transmission, multiplexing, and switching of data  Transmission (Übertragung) – Physical methods to inject data – represented as digital bits and transformed into signals – into a network  Multiplexing (Überlagerung) – Methods to cumulate or crack data onto/from channels – Time, space, or frequency division multiplexing  Switching (Vermittlung) – Methods to deal with formatted data, e.g., packets or cells – Methods to transfer data between entities © 2024 Burkhard Stiller M02-21  Multiplexing  Transmitting more than one signal over one path – Space-division multiplexing More than one physical path is grouped together  Splitting a physical path and occupying the two-dimensional continuum of frequency and time – Frequency-division multiplexing Quantity of information transmitted using a certain range of frequency Example, FM radio – Time-division multiplexing Quantity of information transmitted using a certain range of time  Applying coding techniques onto multiple signals – Code-division multiplexing – Orthogonal codes, mainly applied in wireless communications © 2024 Burkhard Stiller M02-22  Switching (1)  Terminology is widely used with very different meanings!  Handle packets at wire-speed (Layer 1)  Layer-2-Switching – cf. before  Layer-3-Switching – Combination of switching speed and router functionality – Similar terminology: Routing switches or IP switches – Identification for common traffic flows on layer 3 and switch these flows on the hardware level for speed. Other traffic will be routed as usual  Layer-4-Switching – Includes application-level control by applying filters, e.g., security, and QoS-control on specific application flows © 2024 Burkhard Stiller M02-23  Switching (2)  Hubs vs. Switches (Layer 1) – Similar locations in networks – Hubs repeat all packets while switches examine all of them – Switches require address examination and forwarding Store-and-forward: Analyze the entire packet Cut-through: Only examine destination and forward Blocking vs. non-blocking architectures Buffering: backpressure or large buffers BCDE BCDE A Hub F A Switch F to E to E © 2024 Burkhard Stiller M02-24  Topologies and Switching (1)  Backbone  Backbone Switching (collapsed backbone)  Multi-switch Backbone © 2024 Burkhard Stiller M02-25  Topologies and Switching (2)  Workgroup Segmentation (decentralized)  Workgroup Segmentation (centralized)  Micro Segmentation © 2024 Burkhard Stiller M02-26  Switching Techniques – Overview  Switching Techniques – Based on circuits, cells, frames, or packets Simple Intricacy Complex Fixed Behavior of Bandwidth Variable Circuit Multi-rate Fast Cell Fast Frame Frame Packet Switching Circuit Circuit Switching Packet Relay Switching Switching (STM) Switching Switching (ATM) Switching  Circuit switching suffers from fixed bandwidth constraints for bursty traffic  Packet switching allows for variable bandwidth © 2024 Burkhard Stiller M02-27  Circuit Switching (1)  Establishment of a connection (circuit) required  Example comprises the telephone networks or the Narrow-band Integrated Services Digital Network (N-ISDN)  Synchronous Transfer Mode (STM) – Time-division multiplexing-based – Basic unit of transmission is a time slot – Periodic frames contain a fixed number of time slots which are dedicated to specific circuits due to the establishment phase E.g., periodicity of 8 bit every 125 s for 64 kbit/s Static mapping of incoming to outgoing links © 2024 Burkhard Stiller M02-28  Circuit Switching (2)  Circuit information are stored during establishment times in a translation table (switching table)  Bit error rates are caused Incoming Time Outgoing Time Link Slot Link Slot by single bit errors (switch- ing malfunction) or bursts l1 1 O1 3 2 2 O2 (loss of synchronization) … O3 … m 1  Delay is determined by the … O1 4 l2 1 propagation delay and the 2 O2 2 … O3 … processing in switches m … m – Bounded to 450s by ITU-T O1 lm 1 1  Inflexible, e.g., G.703 PCM 2 … O2 m … O3 – Fixed bandwidth m … 1 © 2024 Burkhard Stiller M02-29  Multi-rate and Fast Circuit Switching  Overcoming the inflexibility of fixed bandwidth – Multiple basic fixed rate bandwidth circuits are used E.g., n times m kbit/s – Limits the static bandwidth behavior – Increases the complexity of switching systems – Synchronization between circuits has to be maintained explicitly – Selection of basic bandwidth m is still a problem Solution with multiple (i) basic rates mi feasible  Circuit establishment only, if data are to be sent – Fast allocation of resources on a burst basis in terms of Bandwidth is defined as k times of a basic circuit rate Destination is stored within the switching system Circuit is identified by a header in the signaling circuit On sending data, resources are allocated immediately – Complexity of the switching system still remains © 2024 Burkhard Stiller M02-30  Packet Switching Techniques Packet Switching “Slow” Packet Fast Packet Switching Switching Frame Frame X.25 Relay Switching Variable Size, Variable Size Fixed Size Connection- oriented DQDB ATM Connection- Connectionless oriented © 2024 Burkhard Stiller M02-31  Packet Switching  Based on user data that are encapsulated in packets  Link-based error control in complex protocols  Variable sized packets require a complex buffering  Concept is based on technology available in the 60s – Erroneous links – Low bandwidth links – Software-based protocol implementations  High delays (due to retransmissions) and low speed (due to protocol processing) avoid the support of real-time and multimedia traffic  X.25 is the oldest example of packet switching nets © 2024 Burkhard Stiller M02-32  Fast Packet Switching  Applies to systems operating at high speed switching  Synonyms of this historic name are – Asynchronous Time Division (ATD) or the well-known – Asynchronous Transfer Mode (ATM)  Asynchronous operation between sender and receiver is possible (no clock dependency) – Applying fixed-size packets: cells – Reached via inserting/removing “idle packets” (content-less information) in a flow of data  Main advantage – Transport of data feasible, irrespective of their characteristics, such as bit rate, bursty nature, or quality © 2024 Burkhard Stiller M02-33  Frame Switching and Relaying  Frame switching became feasible due to higher bandwidths available per link – No multiplexing of logical channels required – Traditional functionality is performed on the link level  Frame relaying became feasible due to lower error rates available per link – Less protocol functionality on the link level (no retransmissions, flow control, and multiplexing) required – An error detection function is required only to discard erroneous frames (CRC checksum) © 2024 Burkhard Stiller M02-34  Frame Relay (1)  Frame Relay is a standardized WAN technology specifying the physical and data link layers of digital telecommunications channels using a frame switching methodology  Frame Relay supports connection-oriented services  Subscriber Network Interface (SNI) defined between customer (router) and PTO equipment PTO: Public Telecommunication Operator – Support of pure data, not particularly voice – Multiplexes flows of data being divided in data blocks – Flows are carried in virtual channels which may exceed their bandwidth as other channels are idle – Frame Relay may carry X.25 packets/frames  Performance – Different implementation approaches exist, 2 Mbit/s access speeds are common – Insufficient guarantees on bandwidth and delay variation © 2024 Burkhard Stiller M02-35  Frame Relay (2)  Physical and data link layer specifications available  The data link layer is based on LAPD (ISDN) – Data link services: addressing (DLCI, local significance) – Error control is left out as an end-to-end function – Core LAPD frame Byte 1 2 variable 2 1 Flag Address Payload FCS Flag 01111110 01111110 DLCI C/R EA DLCI FECNBECN DE EA bit 6 1 1 4 1 1 1 1 DLCI: Data Link Connection Identifier FECN: Forward Error Congestion Notification C/R: not used BECN: Backward Error Congestion Notification EA: Extended Address DE: Discard Eligibility © 2024 Burkhard Stiller M02-36  Frame Relay (3)  Frame Relay DLCI assignments 0 Reserved for Call Control Signaling 1-15 Reserved 16-1007 Assigned to Permanent Virtual Circuits (PVC) 1008-1022 Reserved 1023 Local Management Interface  A simple sample Frame Relay network 1007 User A Frame Relay Interface Frame Relay Interface 16 Permanent Virtual Circuits Frame Relay Interface User B 1007 User C 16, 145, 1007: DLCI 145 © 2024 Burkhard Stiller M02-37  Comparison  Summary of important functional differences Frame Frame X.25 Switching Relay Connection-oriented X X X Connectionless – or X – Frame Boundaries X X X Bit Stuffing X X X CRC X X X Error-Control ARQ X X – Flow-Control X X – Multiplexing of log. Channels X – – Heavy weight Light weight Technology CRC: Cyclic Redundancy Check © 2024 Burkhard Stiller M02-38  Characteristics of High Speed Networks  Major characteristics of high speed networks – Low bit error rate (fiber optical media) – Higher packet error/loss rate (buffer overflow) – Existing Jitter (different buffer lengths) – Small transmission units (cells), – Many connections (context data) – High bandwidth (fiber optical media) – Extreme bandwidth-delay product  Protocols have to deal with these issues to alleviate influences of delayed, corrupted, or lost data  Protocols have to be aware of significant network improvements to reward multimedia applications © 2024 Burkhard Stiller M02-39  Effects on Data in High Speed Transit  Buffer overflows dominate occurring errors – The effect grows worse for fast and real-time traffic  Error recovery has to be a trade-off between a waste of bandwidth or extending delays – Bandwidth is much cheaper than tolerating high delays – Multimedia applications normally do not like delays  Signal delay is dominating the transmission delay – This grows worse for high transmission rates  A huge amount of data is in transit within high speed and long distance networks  Path Capacity PC: PC = B * Dsignal B: Bandwidth, Dsignal: Signal delay. © 2024 Burkhard Stiller M02-40  File Transfer Example File size: 1 Mbyte Link: San Diego – Boston BOS Signal delay: 25 ms SAN 5000 km  64 kbit/s channel – 64 kbit/s * 25 ms = 1,600 bit – 0.02 % of file on link SAN BOS  2 Mbit/s channel – 2 Mbit/s * 25 ms = 50,000 bit 1,600 bit – 0.6 % of file on link  1 Gbit/s channel 50,000 bit – 1 Gbit/s * 25 ms = 25,000,000 bit – 8 ms for transmission, 17 ms idle 25,000,000 bit Bits are smaller – not faster ! © 2024 Burkhard Stiller M02-41  Path Capacity Calculations – Link length: 1,000 km – Optimal physical delay per 1 km: 5 s – Transmission delay due to signal delay: 10 ms for the link – PC = B * Dsignal: PC = 10 kbit/s * 1,000 km * 5 s/km = 50 bit Type of Network Low Capacity High Capacity Transmission Rate 10 kbit/s 1 Gbit/s Path Capacity: Bits on Link 50 bit 5 Mbit Delay per Bit 100 s 1 ns Delay per 1,000 bit 100 ms 1 s – Round-trip time (RTT): > 20 ms – Time for 3-way hand-shake: > 30 ms – 10 Mbit ≈ 157 packets á 8 kByte ≈ 26,000 cells á 53 Byte © 2024 Burkhard Stiller M02-42