3 IOM - iDirect DataComm, 061407

Copyright NoticeiDirect Technologies Technical Training ManualCopyright 2002; 2004 - 2007, iDirect, Inc. All rights reserved. This training material may not be reproduced, in part or in whole, without the permission of iDirect, Inc. All other brands or product names are trademarks or registered trademarks of their respective holders.Printed in the USA.No part of this work covered by copyright may be reproduced in any form. Reproduction, adaptation, or translation without prior written permission is prohibited, except as allowed under the copyright laws.This publication is provided by iDirect Technologies as-is without warranty of any kind, either express or implied, including, but not limited to, the implied warranties or conditions of merchantability or fitness for a particular purpose. iDirect Technologies shall not be liable for any errors or omissions which may occur in this publication, nor for incidental or consequential damages of any kind resulting from the furnishing, performance, or use of this publication.Information published here is current or planned as of the date of publication of this document. Because we are improving and adding features to our products continuously, the information in this document is subject to change without notice.

IOM Chapter 3 - DataCom Basics

Datacom Training ObjectivesDescribe different types of data trafficReal Time vs. non-Real TimeCharacteristics & processing different traffic typesDiscuss Packet Encapsulation and handling techniquesIdentify iDirect frame and packet structure formatsDescribe various optional iDirect efficiency features availableSegmentation and Reassembly (SAR)Timeslot FeatheringFrequency Hopping vs. Carrier Grooming, etc.Forward Error Correction (FEC)Helps to provide error free delivery of data Discuss many important DataCom terms & principalsIP, TCP, & UDP protocolsPacket Header format and breakdownTCP Acceleration, or SpoofingQuality of Service (QoS) methodologyTraffic Engineering as it relates to QoS


Traffic Types ComparisonVoice/Video Traffic CharacteristicsReal Time Protocol (RTP) ApplicationsTime sensitive informationSensitive to Delay and Delay Variation (Jitter)Deliver information in real-time or not at allInformation content directly affected by delay (time)Not sensitive to bit errors (uncompressed)Information never retransmitted Multi-media Applications & Image Processing Data Traffic CharacteristicsNot Real-time High Speed Data (very Error sensitive)Not sensitive to Delay or Delay VariationSensitive to even a single bit error (retransmissions)Information content unaffected by time (delay)Very sensitive to bit errors Information retransmitted on bit errorLarge file transfer (ftp, etc.)


Generic OSI ModelAPPLICATION LayerPRESENTATIONSESSIONTRANSPORTNETWORKData Link LayerMAC SublayerLLC SublayerPhysicalOpen Source Interconnection (OSI) Model/Protocol StackPerforms common application services and supports end-user processes, (Telnet, FTP, e-mail, etc.).Provides services to the Application layer for syntactical differences in data representation within the end user systems.Manages interaction between end-user processes. Establishes check-pointing, adjournment, termination and restart procedures.Provides transparent transfer of data between end users. It ensures the method for accomplishing complete data transfer.Provides the functional & procedural means of transferring variable length data sequences from a source to a destination while maintaining the quality of service requested by the Transport layer. Performs network routing & error control functions. Provides the functional and procedural means to transfer data between network entities. Includes Media Access Control (MAC) layer, which controls network access & Logical Link Control (LLC) layer for frame sync, flow control and error checking.Establishes & terminates connection to a communications medium. It is the hardware layer providing physical means for sending and receiving data.Layer7654321TCP/UDPIPEthernetHTMLHTTPMMF, SMFUTP, STP, CoaxRF FrequencyPhysicalMedia


Ethernet Packet EncapsulationPacket (Application) DataTCP Packet(Layer 4)Ethernet Packet (Layer 2)TCP Packet Information FieldEthernet Packet Information FieldIP Packet Information FieldIP Packet (Layer 3)IP HeaderProtocol Header *TCP, UDP, ICMP, etc.Ethernet HeaderLayer 5* Header = Overheard Data, not part of user data traffic added to the head of the packetLayer 4CRCLayer 2Layer 3


Communications Concepts - SAR Dequeue22345ij1RF OutQoS DistributorVoIP Packets (70 Byte)FTP Packets (1500 Byte)abcdefgh1The Large FTP Packet may delay voice, other Real Time Protocol (RTP) in a congested stateAs packets are divided into equal sized packets, a voice packet has to wait a maximum of one burst.Packet Arrival TimeSAR not enabledSAR enabled (affixes small, 2 byte SAR header to each segment maintains QoS for each segment)Generic Ethernet Packet Structure consideredSAR Segmentation and Reassembly1a. . . . . . .nab2cbcSAR


Forward Error Correction (FEC) Overhead & Rate ConversionsThe lower the selected FEC rate, the greater the number of FEC bits that will be required to support that rateA lower rate requires a greater allowance for FEC overheadA lower rate provides better data integrity, albeit at the cost of available bandwidthA smaller VSAT reflector may be supported at the remote using lower FEC ratesConsequently, a greater number of bits for FEC overhead reduces the user traffic payload in the affected data streamTerminology - Data Rate ConversionsIP User Rate + iDirect Overhead = Info Rate(TX) Info Rate / FEC Rate = Transmission Rate; (RX) Info Rate x FEC = Receive RateTransmission or Receive Rate = Symbol Rate (BPSK)Transmission or Receive Rate / 2 = Symbol Rate (QPSK)Symbol Rate x 2 = Transmission or Receive Rate (QPSK)Symbol Rate x Channel Spacing = Allocated BW


Anatomy of A Satellite Space Segment* - Example Case Study, 1.25 Channel Spacing


Protocol Structure Outroute


Balloon callout. Select shape and start typing. Resize box to desired dimensions. Move control handle to aim pointer at speaker.

Payload

SOF

FEC

404

106

Protocol Structure Inroute



43

1

8

2

6

2

2

73

1

IP Payload

GB

Unique Word

DH

Link Layer

PAD

CRC

FEC

GB

56

60

72

44

394

106

Star TDMA vs. Mesh TDMA 2 Additional bytes required for Mesh header



IP Payload

GB

Unique Word

DH

Link Layer

PAD

CRC

FEC

GB

72

44

Inroute Frame Structure



Time Slot Assignment - CGCarrier Grooming Inroute GroupOne Inroute Frame (125 msec Typical)Allocated Time Slots per Remote


Time Slot Assignment - FHFrequency Hopping Inroute GroupOne Inroute Frame (125 msec Typical)Allocated Time Slots per Remote


Feathering Timeslot AllocationSignificant variation in delay causes jitter.Bandwidth allocated to R1


Scheduled Dedicated TimeslotWith the default configuration every remote is given a dedicated timeslot, in every frame. For ex., 5 Remotes using 5 Timeslots, with each remote getting a timeslot every frame.If a remote is configured to have its dedicated time slot guaranteed once every 2 frames, then 10 remotes will need only 5 timeslots (maximum setting is one time slot every two seconds)This allows one to oversubscribe an inroute/inroute group at a much higher ratioSome example applications would be business continuity and low bandwidth networks that need a guaranteed amount of bandwidth


iDirect Hub Logic/SubnetsUpstream Router is the gateway device for all network throughputNetwork (User) TrafficM&C (NMS) TrafficTwo subnets assigned for hub componentsUpstream & TunnelTypically full class C for each; minimum .248 subnet maskUpstream subnet = network trafficTunnel subnet = Primary M&C trafficOnly Hub Line Cards are IPd with Tunnel subnet addressHLC default Gateway = Upstream Router Tunnel Interface


PP Blade 2Eth0 IP: 192.168.8.6Eth1 IP: 192.168.8.67

NMS BackupEth0 IP: 192.168.8.4

Upstream ROUTER

IOM Team 8 Logic Diagram

Eth0 interface IP: 192.168.8.1 Eth1 interface IP: 192.168.8.65

IOM Chassis S/N: N/A , Sim. IP Addr: 192.168.8.2

Slot 13 Tx/Rx ID: M1D1-T.40700 IP: 192.168.8.68 Slot 14 Rx only ID: M0D1.40516 IP: 192.168.8.69Slot 15 Stdby ID: M1D1-T.13091 IP: 192.168.8.70

PP Blade 1Eth0 IP: 192.168.8.5Eth1 IP: 192.168.8.66

NMS OnlineEth0 IP: 192.168.8.3

Access via Table in Training Room A via Gray CAT5

Utility (NMS GUI) PCEth0 IP: 192.168.0.dhcp

Internet Protocol (IP)IP Protocol Connectionless ProtocolSpecifies only Best Effort ; provides unreliable packet deliveryNo retransmission of IP packetsPackets discarded if network resources are insufficientSource & Destination IP Facilities do not handshakeSpecifies the format of all data IP Software performs the routing functionPackets are treated independentlyPackets may take different paths through the networkIP Provides packet delivery service to Transport Layer protocolsIP provides common, consistent, universal addressing techniqueIP defines set of rules that embody packet transmission & deliveryIP rules:Specify how routers should process packets (Routing, ToS, precedence, fragmentation)Specify when and how to generate error messages (ICMP)Conditions that govern discard and/or duplication of packets (multicast)


Transmission Control Protocol (TCP)TCP is a Connection Oriented protocolSource and Destination MUST agree to the transmission and reception of information PRIOR to the transfer of user trafficDestination must agree to receive the informationSimilar to a standard telephone callAll transmissions are acknowledged Provides guaranteed end to end deliveryDictates procedures to agree when transfer is completeSpecifies the format of the information Specifies acknowledgements that the information was receivedSpecifies method to ensure information was received correctlyRetransmission IS REQUIRED if acknowledgement not receivedin the event of a lost or corrupted packetDetermines how a machine distinguishes between multiple destinationsProvides recovery from errors, or lost & duplicated packetsResponsible for Flow ControlTCP & IP operate over dial-up, LAN, Optic, high & low speed WANsSatellite links (if inherent of round-trip latency can be overcome)Successfully done when required acknowledgements are spoofed


Transmission Control Protocol (TCP)TCP Multiplexes & Demultiplexes data to/from applicationsMust be able to distinguish data flows between destinationsTCP uses Port IDs & destination IP address to distinguish flowA TCP Port is a queue into which TCP protocol places data-gramsTCP uses connection abstractions such as:Source/Destination Port(s)Host Address:Port and/or Source Address:Port pairingsSource (65.168.20.1:100) Destination (10.10.200.1:200)Source and Destination pairing to identify a data flowRequires only one local port to accommodate many data flows for many local applicationsTCP (and UDP) encapsulates the data traffic, or IP PacketIP Packets are the delivered packets (payload)IP Packets are the single packet that traverses the networkIn a routed network Layer 2 packets live only point-to-pointEthernet Packets live only between adjacent ports


Transmission Control Protocol (TCP)SEND SYNCSYNC bit: SetSeq Num: AsnACK bit: NOT SetRECV SYNCSYNC bit: SetSeq Num: AsnACK bit: NOT SetSEND SYNCSYNC bit: Set Seq Num: BsnACK: Asn+1RECV SYNCSYNC bit: SetSeq Num: BsnACK: Asn+1SEND ACKSeq Num: Asn1ACK: Bsn+1RECV ACKSeq Num: Asn1ACK: Bsn+1TCP Connection Establishment


User Datagram ProtocolUser Datagram Protocol (UDP) Connectionless ServiceBest Effort packet delivery serviceHow good is best?Not guaranteed deliveryNo acknowledgements ever providedSource sends information without respect to agreement by any destination to receive the informationSimilar to announcements over a public address systemIf not received, missing transmissions are never resent


TCP AccelerationTCP Acceleration, or Spoofing definedTCP requires destination to acknowledge receipt of every packetThis can result in delay when Satellite round-trip LATENCY (~500msec) is introduced into the data pathiDirect solution acknowledges TCP packet receipt for the machine (or Host) that requested the informationRemote site thinks the iDirect customer has received packets and sends more dataSometimes referred to as SpoofingiDirect remote tells Hub if it did not get a data packetFaster data transmissionsIncreases overall data throughputMaximizes available bandwidth


Impact of Satellite Delay~530 msecRound TripLatencyTCP/Web Acceleration


TCP Spoofed (Accelerated)Minimal LatencyTCP/Web Acceleration


Quality of Service (QoS)Quality of Service (QoS) Generically DefinedProviding CONSISTENT delivery of traffic across a networkQoS effort is to impart delay on data flows that can tolerate delay Reduce delay on those flows that can not tolerate delayQoS is required due to different traffic characteristics such asReal-time applications (VoIP, streaming video/audio)Real-time DATA applicationsnon Real-time data applicationsPriority users - QoS applied to a connection rather than a packetApplies to all packets in that connection (data flow)QoS is implemented by granting first (or early) access to the TRANSMIT Queue using one of three optional/configurable QoS methodsPriority QueuingClass Based Weighted Fair QueuingBest Effort QueuingType of Service (TOS) applies to specific packet (in IP header field)


Quality of ServiceQoS ImplementationsTypically required to support Service Level Agreements (SLA)Provides ability to distinguish between different traffic typesDesigned to enable one user to be favored over anotherImplemented through the management of resources such asBandwidth allocationPacket Loss, reliabilityLatency and jitterWill not eliminate or reduce network congestionA heavily over-subscribed network may simply require more bandwidth to be allocatedQoS capabilities will not improve service if host is overloaded


Quality of ServiceQoS at a higher levelApplication QoS - Realtime classification & prioritization of trafficTraffic prioritization can be performed usingSource and/or Destination IP AddressSource and/or Destination IP SubnetSource and/or Destination Port NumberSource and/or Destination Port RangeDiffserv and ToS BitsVLAN IDProtocol (TCP, UDP, HTTP, ICMP, IGMP)Configurable Queue Lengths For Each Service LevelA percentage of bandwidth is assigned to each service levelConfigurable drop policyTraffic can be rate limited in both directions


Network QoSNetwork QoS DescriptionHierarchical Prioritization CapabilityMost Flexible QoS Capability in IndustryMost Comprehensive Prioritization CapabilityNetwork QoS - With CIR guarantees service at network levelPrioritize real-time traffic over non-real-time traffic across shared satellite bandwidthAllocates traffic demand for each remote appropriatelyPrevents non real time traffic from one remote impactingReal Time Traffic on a different and unrelated remote


Quality/Class/Type of ServiceDifferent Service ClassificationsQuality of ServiceCapability to provide delivery of information across a network connection with consistent performance characteristicsClass of ServiceThe ability to establish categories of QoS (Gold, Silver, Bronze)A method of specifying and grouping applications and traffic into QoS categoriesType of ServiceSometimes referred to as IP PrecedenceGenerally unused, but important QoS setting for iDirect Network Accelerator Provides the ability to indicate to network devices that this packet should be afforded highly reliable transport (no drops) or low delay


Traffic EngineeringTraffic Engineering as a QoS Management ToolTraffic Engineering is the umbrella terminology of network QoSDesigned to optimize the operation of communications networksTraffic Engineering is composed of:Measurement (post-deployment)Reporting (before, during & after deployment)Control (adjusting configuration parameters to optimize network)Traffic Engineering is classified as:Traffic OrientedQuality of Service related issuesPacket Loss MinimizationDelay MinimizationThroughput MaximizationEnforcement of Service Level AgreementsResource OrientedEnsure areas of Network are properly utilizedLimit the Under / Over Utilization of network areasOptimum Bandwidth UtilizationCongestion Control


Traffic EngineeringTraffic Engineering RevealedCongestion Control - most important objective of Traffic EngineeringPermits enforcement of service level agreementsTraffic Engineering is designed to limit long duration CongestionTransient Congestion not addressedPrimary causes of network congestionNetwork resources insufficient to handle the loadTraffic streams are inefficiently mapped to available resourcesInsufficient network resources addressed byCapacity ExpansionFlow Control TechniquesRate LimitingRouter Queue ManagementWindow Flow ControlEfficient mapping of traffic flows addressed by Traffic Engineering Architecture and load balancing policiesTraffic Engineering policies are implemented through other specificationsReservation Protocol (RSVP)Differentiated Services (Diff-Serv)


Chapter 3

Installation, Operations & Maintenance (IOM)

iDS v7.0 June 2007Data Communications Basics

Thanks, . . .

Questions? . . .


Requisite Copyright noticeObjective for this chapter/moduleDescription of Real vs Non-Real Time data traffic.Generic OSI model (for reference only really)Packet encapsulation from session layer (layer 5) with headers added thru each layer to the Ethernet-Link layer, layer 2. (Animated if you turn it on . . . )This is the Segmentation and Reassembly (SAR) description. The above example describes the process of segmenting larger data packets into smaller chunks. The segments are then interspersed among equal, or smaller sized VoIP (real-time) packets allowing the smaller sized packets access to the distributor queue. If not enabled, the larger packets would delay the smaller packets from being processed until its full transfer is completed. This would result in unwanted delay, or jitter, resulting in potentially degraded performance.Additional information for the implementation of this optional feature can be found by turning to the description in Appendix A.4.5 on page 163-164, and/or to Chapter 7, Configuring Remotes, Up/Downstream SAR on page 109 for the configuration details in the iDirect Network Management System (NMS) iBuilder users Guide, dated April 6, 2005.SAR can be enabled for both the Downstream and Upstream traffic flows independently and, when enabled, the SAR segment size can be entered default segment size is 70 bytes.SAR default is not-enabled on the Downstream traffic, but it is defaulted as enabled in the Upstream direction.

Set-up for showing rate conversion example (coming up).Pictorial overlaying all rates just determined.

DS = DownstreamUS = Upstream Generic OSI model (for reference only really)Our typical generic logic diagram. . . Upstream LAN is the traffic LAN. . . All hub traffic uses this Upstream LAN as the default, except for the HLCs which are the only hub component to have an address on the Tunnel LAN segment (as the sole IP address). PP also has second NIC address on Tunnel, but Upstream LAN and eth0 interface are configured as PP default gateway and gateway device (coming up in next chapter).Network traffic does actually process through the tunnel port connection between the HLC and PP, giving the Tunnel LAN segment its name. However, the data traverses the PP from eth1 to eth0 and exits via the previously mentioned Upstream/eth0 port as the configured default.Tunnel LAN is also known as the Monitor and Control LAN since the NMS uses the PP Tunnel LAN address to query the PP and gather status and statistics for network operation. PP always replies on the Upstream LAN, as always.Our typical generic logic diagram. . . Upstream LAN is the traffic LAN. . . All hub traffic uses this Upstream LAN as the default, except for the HLCs which are the only hub component to have an address on the Tunnel LAN segment (as the sole IP address). PP also has second NIC address on Tunnel, but Upstream LAN and eth0 interface are configured as PP default gateway and gateway device (coming up in next chapter).Network traffic does actually process through the tunnel port connection between the HLC and PP, giving the Tunnel LAN segment its name. However, the data traverses the PP from eth1 to eth0 and exits via the previously mentioned Upstream/eth0 port as the configured default.Tunnel LAN is also known as the Monitor and Control LAN since the NMS uses the PP Tunnel LAN address to query the PP and gather status and statistics for network operation. PP always replies on the Upstream LAN, as always.Our typical generic logic diagram. . . Upstream LAN is the traffic LAN. . . All hub traffic uses this Upstream LAN as the default, except for the HLCs which are the only hub component to have an address on the Tunnel LAN segment (as the sole IP address). PP also has second NIC address on Tunnel, but Upstream LAN and eth0 interface are configured as PP default gateway and gateway device (coming up in next chapter).Network traffic does actually process through the tunnel port connection between the HLC and PP, giving the Tunnel LAN segment its name. However, the data traverses the PP from eth1 to eth0 and exits via the previously mentioned Upstream/eth0 port as the configured default.Tunnel LAN is also known as the Monitor and Control LAN since the NMS uses the PP Tunnel LAN address to query the PP and gather status and statistics for network operation. PP always replies on the Upstream LAN, as always.Our typical generic logic diagram. . . Upstream LAN is the traffic LAN. . . All hub traffic uses this Upstream LAN as the default, except for the HLCs which are the only hub component to have an address on the Tunnel LAN segment (as the sole IP address). PP also has second NIC address on Tunnel, but Upstream LAN and eth0 interface are configured as PP default gateway and gateway device (coming up in next chapter).Network traffic does actually process through the tunnel port connection between the HLC and PP, giving the Tunnel LAN segment its name. However, the data traverses the PP from eth1 to eth0 and exits via the previously mentioned Upstream/eth0 port as the configured default.Tunnel LAN is also known as the Monitor and Control LAN since the NMS uses the PP Tunnel LAN address to query the PP and gather status and statistics for network operation. PP always replies on the Upstream LAN, as always.Our typical generic logic diagram. . . Upstream LAN is the traffic LAN. . . All hub traffic uses this Upstream LAN as the default, except for the HLCs which are the only hub component to have an address on the Tunnel LAN segment (as the sole IP address). PP also has second NIC address on Tunnel, but Upstream LAN and eth0 interface are configured as PP default gateway and gateway device (coming up in next chapter).Network traffic does actually process through the tunnel port connection between the HLC and PP, giving the Tunnel LAN segment its name. However, the data traverses the PP from eth1 to eth0 and exits via the previously mentioned Upstream/eth0 port as the configured default.Tunnel LAN is also known as the Monitor and Control LAN since the NMS uses the PP Tunnel LAN address to query the PP and gather status and statistics for network operation. PP always replies on the Upstream LAN, as always.TDMA Time Slot Feathering is discussed. This process is done automatically and is not an option-able feature. Additional information covering the details for this always on feature can be found by turning to the description in Appendix A.4.5 on page 163 in the iDirect Network Management System (NMS) iBuilder users Guide, dated April 6, 2005.Discussion of the concept of scheduled dedicated time slots. Top example shows default with each remote receiving one time slot per frame, every frame.Bottom half gives an example of dedicating one time slot, per remote, every other (or every second) frame. If using 125 ms frame length (typical) the maximum setting for the dedicated time slot of 2 seconds would result in one slot every 16th frame.This is only for idle time ICMP replies, UCP info, etc. If more BW is required or demanded then the remote gets whatever the PP can allocate from the time plan using existing QoS profiles.Opens discussion on the Scheduled Dedicated Time Slot technique.Our typical generic logic diagram. . . Upstream LAN is the traffic LAN. . . All hub traffic uses this Upstream LAN as the default, except for the HLCs which are the only hub component to have an address on the Tunnel LAN segment (as the sole IP address). PP also has second NIC address on Tunnel, but Upstream LAN and eth0 interface are configured as PP default gateway and gateway device (coming up in next chapter).Network traffic does actually process through the tunnel port connection between the HLC and PP, giving the Tunnel LAN segment its name. However, the data traverses the PP from eth1 to eth0 and exits via the previously mentioned Upstream/eth0 port as the configured default.Tunnel LAN is also known as the Monitor and Control LAN since the NMS uses the PP Tunnel LAN address to query the PP and gather status and statistics for network operation. PP always replies on the Upstream LAN, as always.Description of IP protocol . . . TCP Description . . . TCP Description continues . . .TCP Description continues . . .Basic acknowledgements required for each transaction . . . UDP Description . . . TCP acceleration also called spoofing try to discourage the use of this term

Key differentiator in iDirect solution

Available because of the tight integration with Hub NetModem and NetModem II/II+ products, both pieces are essential to providing the iDirect solutionQoS definitions/description . . .

ToS is important to point out here since our NA uses the ToS bit for the encrypted data TCP Acceleration.Further QoS topics . . . QoS detail continued . . . Describing QoS Service Level criteria which will be coming up later. . . (NMS/iBuilder chapter) . . . Applying QoS at the Network Level. . . .More QoS detail . . . Traffic Engineering . . . Part of QoS concept . . .Traffic Engineering . . . Congestion . . . Part of QoS concept . . .

3 IOM - iDirect DataComm, 061407

Documents