SOFTWARE DEFINED RADIO WIDEBAND NETWORKING QOS TEST BED

Greg Osborn', Frank James', Becky Swinford', Al Stewart2, and Scott Chuprun'General Dynamics C4 Systems (first.lastggdc4s.com)

IScottsdale, AZ, 2Taunton MA

ABSTRACT

The future networked architectures currently in develop-ment represent a significant advancement from today'sdeployed military communications infrastructure. Giventhe substantial effort in developing andfielding these newtechnologies, transitional test beds are needed to evaluatenetworking objectives against simulated results. This pa-per describes a flexible test bed solution that supports theevaluation of communication protocol layers 1-3 in thecontext ofa multi-channel software defined radio. The testbed provides a path for evaluating emerging high datarate waveforms (e.g. tactical OFDM), adaptive cross-layered Quality of Service behaviors, and network topol-ogy management using the Distributed Network Agent(DNA) tool. It also demonstrates the feasibility of inte-grating networking technology into fielded SDR platforms.Both the current and planned capabilities of the test bedare described.

INTRODUCTION

The scope of this paper is the description of a network testbed that supports the end to end (E2E) analysis of IP traf-fic protocols through High Assurance IP Encryptor,(HAIPE) devices and across a black (i.e. unsecured) pri-vate network comprised of: switches, routers, local andwide area wireless links, consistent with the proposed fu-ture defense network architectures. The test bed is particu-larly focused on demonstrating the interoperability ofcommercial devices and fielded software defined radio(SDR) technology to achieve quality of service (QoS) ob-jectives in the physical, link and network layers acrosssmall (<25 node) wireless radio networks. The paper isorganized as follows:* First section of the paper introduces the latest military

network architectures and summarizes the importantcommon objectives and technologies

* Second section describes the goals and objectives ofthe desired network test bed

* Third section describes the architecture, hardware andsoftware components comprising the test bed

* Fourth section summarizes progressive developmentof protocol layers 1-3 in the SDR

* Fifth section provides a summary of the goals and ob-jectives of the SDR test bed

It is hoped that this test bed will identify protocol relatedissues and further our understanding of the proposed no-tional architectures. The research focuses on verificationthrough implementation. While simulation of larger net-works is a valuable approach, it is not within the scope ofthis investigation to assess the scalability of the proposedfuture network architectures. It is intended to provide atransition platform and network test bed on which toevaluate a subset of the protocols and techniques, with anemphasis on QoS, as required by the armed forces in theproposed future architectures.

MILITARY NETWORK ARCHITECTURES

The Department of Defense's (DoD) transformational vi-sion is one of seamless global connectivity through crossechelon communication capabilities and full-spectrumdominance. Joint Vision 2020 provides the conceptualtemplate for the DoD's vision and is comprised of severalkey programs [1]. The key military network architecturesin development today are described below and depicted inFigure 1.

Global 1norP Gnd(GIG)

Tranffmati

X- - --r

p~~~~l-nal mm1(TC)

X-_ I!,.... ,

t ................. u~~~~~~~

us

G &E

Joint Vision 2020 is Fulfilled as GIG Merges FCS and ForceNet

Figure 1 - Joint Vision 2020

(C 2005 General Dynamics. All rights reserved.

FCS

I of 7

The Global Information Grid (GIG), described in [2],represents a new paradigm in global communications, in-formation management and security. The goal of the GIGis to provide a seamless, secure and globally intercon-nected communications system. The GIG is comprised offour conceptual transport tiers. Each tier is comprised ofvarious technologies that are tasked to provide seamlessend-to-end communication, plug and play interoperability,security, and high capacity bandwidth on demand.

The Army's Future Combat System (FCS), a major com-ponent of the GIG, transformation program runs across allthe military services [3]. FCS, using advanced communi-cation systems, links 18 individual systems in addition tothe network and the warfighter to form one large "systemof systems". The Warfighter Information Network-Tactical (WIN-T) will serve as the backbone and manage-ment of the FCS.

WIN-T Network Architecture

The WIN-T network [4] is the Army's tactical communi-cation network providing command, control, communica-tions, computers, intelligence, surveillance, and reconnais-sance (C4ISR) capabilities. WIN-T enablescommunications at all echelons and provides the capabilityof the simultaneous communication of voice, data andvideo. WIN-T leverages the capabilities of the Joint Tacti-cal Radio System (JTRS) and wireless local area network(WLAN) and wideband radio capabilities.

JTRS Initiative

JTRS is a DoD transformational program and a part of theGIG that is key to achieving informational superiority asenvisioned in Joint Vision 2020. The Joint Tactical RadioSystem (JTRS) is a group of common, software-definedradios built around the Software Communications Archi-tecture (SCA) that are interoperable and scalable. TheJTRS family of radios provides wireless, mobile, Line-of-Sight (LOS), and Beyond Line-of-Sight (BLOS) C41 capa-bilities.

ADNS Network Architecture

The Navy's Automated Digital Network System (ADNS)[5] is responsible for the transport of all Wide Area Net-work (WAN) IP services between afloat units and shoresites. ADNS is engaged in converging all voice, video,and data communications between ships and shore to an IPmedium taking full advantage of all RF means aboard shipto transmit data efficiently.

ShipClassified/Unclassified Network

HAI PE A E||HI E| |HIP

Classified/Unclassified Network Classified/Unclassified NetworkShip Ship

ADNS HAIPE Encapsulation Across Black Private Network

Figure 2 - ADNS Notional Architecture Subset

Specifically, ADNS automates the routing and switchingof tactical and strategic C41 data via Internet Protocol (IP)networks linking deployed battle group units with eachother and with the Defense Information Systems Network(DISN) ashore. ADNS uses Commercial Off-the-Shelf(COTS) and Non-Developmental Item (NDI) Joint Tacti-cal Architecture (JTA) - compliant hardware (routers,processors, and switches), and commercial-compliant soft-ware in a standardized, scalable, shock-qualified rack de-sign.

The similarities between the WIN-T WAN routing andswitching architecture being developed for the Army andthe ADNS architecture defined by the Navy lead to a con-solidated communications networks approach for multiplemarket convergence. The similarities extend both to thephilosophy and physical representation. They are to:* Converge all voice, video and data communications to

an IP medium using QoS to honor priorities among us-ers and applications

* Employ HAIPE type devices to isolate and separatelyencrypt the various classification levels

* Employ COTS hardware (e.g. routers, switches, proc-essors) and commercial compliant software

* Strive to promote the efficient use of available satelliteand line of sight communications bandwidth

* Improve overall reliability and flexibility

SDR NETWORK TEST BED

With the background of the previous section in mind, theobjectives of the network test bed are to 1) Demonstrateintegration of new IP based waveforms into an SDR, 2)Implement QoS into the SDR consistent with domain widePer Hop Behavior (PHB) treatment of IP packets, 3)Evaluate techniques for managing networked radio links toachieve robust network topology, 4) Evaluate protocolsand approaches for running multi-hop network traffic

(C 2005 General Dynamics. All rights reserved. 2 of 7

across a wideband waveform, and 5) Explore core to edgesolutions on a small scale. Pursuing these goals enablesthe test bed to model and participate both in existing net-work topologies and in the evolutions toward the emergingfuture network architectures.

The First objective is met by developing a full duplex2.4Mbps IP networked OFDM waveform running on awideband version of the General Dynamics C4 SystemsSoftware Defined Radio (SDR) platform. This waveformis referred to as the Tactical Wideband Wireless Network(T-WWiN) waveform and is described in more detail in[6]. The SDR platform and waveform allow for experi-mentation in the implementation and configuration of awide variety of communication protocols in layers 1-3.The second objective is met by implementing cross-layered QoS in the physical, link and network layers of theSDR. The SDR provides a unique opportunity to leveragephysical radio link metrics to influence network routingdecisions. The third objective is met by augmenting thetest bed with a network operations solution that learns theradio network topology and either suggests or implementslink changes to improve performance and reliability of apartial mesh topology formed by RF links between SDRsand legacy platforms. The fourth objective is met by im-plementing a layer 2 switched network and layer 3 routingcapabilities and by exploring the relative performance ofthese two approaches. The fifth objective is met by com-bining network devices and protocols similar to thoseplanned for ADNS and WIN-T to achieve end to endHAIPE protected encapsulation across a black privatenetwork comprised of switches, routers and radio commu-nication links.

The test bed capability is intended to grow in two dimen-sions: E2E scope and E2E embedment. Figure 3 andFigure 4 show the growth path in these two areas.

HAIPE and routing into the SDR for tactical environ-ments.

Network Devices Key: E EndPt H HAIPE R Router S Switch W RF

IV H -W R `H ~-E

WR `H -E

V E

VI E WSRH E

SDR Test Bed Progressively Embeds QoS & Network Function

Figure 4 - Test Bed Phase 4-6

The basic test bed hardware configuration is shown inFigure 5. In this picture, voice and data terminals are con-nected to inline network encryption (INE) devices runningthe HAIPE protocol. The INE's are connected to routerswhich are connected to the SDRs via Ethernet ports. De-pending on the test configuration, the SDR either bridgesor routes to other SDRs in the test bed or to its Ethernetinterface.

SDR Wireless Test Bed Routes IP Based Voice and Data

Figure 5 -SDR Internetworking Test Bed

Network Devices Key: E EndPt H HAIPE R Router S Switch A RF

11

IIIE<K4KRiI wL4Y-fR H E

SDR Test Bed Progressively Extends End to End Network Function

Figure 3 - Test Bed Phase 1-3

Phases 1- 3, shown in Figure 3, are completed and demon-strate a secure end to end HAIPE tunnel through a wide-band wireless link. Referring to Figure 4, phase 4 intro-duces a MAC layer for multiple access and multi-hopnetworking. Phase 5 and 6 illustrate the embedment of

The test bed is comprised of three or more General Dy-namics C4 Systems software defined radios (SDR) capableof running the T-WWiN waveform including data link andnetwork layer protocols. The SDR, which is more fullydescribed in [7], includes both digital and analog inter-faces, embedded cryptographic devices, modulation anddemodulation (modem) functionality, and Radio Fre-quency (RF) translation devices. The crypto module sup-ports COMSEC and TRANSEC functions needed for radiolink security. In addition it supports a growth path for em-bedded High Assurance IP Encryption (HAIPE). EachSDR includes four independent full-duplex RF channels.Each RF channel can be configured through a Human Ma-chine Interface (HMI) to operate as a distinct radio type.SDR is designed to be a multi-mode radio that is capable

(C 2005 General Dynamics. All rights reserved. 3 of 7

of being reconfigured within seconds to minutes to meetthe particular needs of the user and the mission at hand. Anumber of legacy waveforms have been adapted for use onthe SDR radio. These waveforms include High-Data-RateTactical OFDM, SINCGARS ESIP, Have Quick II, UHFSATCOM 181, 182, 183 "DAMA", HF-ISB ALE, Link11/TADIL-A, STANAG 4529 (HF NB Modem), VHF-FM, VHF-AM ATC, VHF/UHF-FM LMR, UHF-AM/FM/PSK, and Link-4A/TADIL-C.

acquisition while concentrating on channel compensationtechniques to mitigate multi-path degradations associatedwith maritime operation.

4-QAM, w/o FEC:16-QAM, w/o FEC:

imples: 1024 66*271 @lMspsTime: -lms ~-18ms

Data Bits: 66*1 72*2 41.2MbpsData Bits: 66*1 72*4 4 2.4Mbps

Figure 6 depicts the mapping of network stack functionaity to the primary hardware modules in the SDR for a siigle channel. Note that the embedment of HAIPE is a fiture activity that is not essential for test bed operatioiThe protocol layers are described in more detail in latesections of this paper.

7_ Hardware Module Mapping

ul)0 -~eo) F

I I ITactical EndPoint Only Black Private Network Wideband Wireless Node

Figure 6 -SDR Internal Network Stack

NETWORK FORMATION AND MOBILITY

This section describes the techniques used to build a nework test bed from an existing SDR device and a high dalrate LOS OFDM waveform to achieve IP networked Trates in mobile environments. The development of the Inetworking is accomplished in a phased approach. Initiefforts are focused on low mobility networks. As the telbed matures, it will grow in its ability to handle high(mobility rates. Each progressive development is describein a separate subsection below.

Physical Layer: T-WWiN

The creation of the robust high data rate physical layer ]WWiN waveform for use in SDR IP networking was ch(sen as being representative of the Navy's need for a conmunications waveform to support collaborative missicplanning among multiple ships while involved in Sea Ba:ing and Sea Strike maneuvers. A software communicatiorwaveform was developed that includes fast and reliab]

[1-

11-

Figure 7 - T-WWiN Framing

* Throughput data rates up to 2.4Mbps are achieved on aer single SDR channel based on the frame format and timing

shown in Figure 7. As currently implemented, each physi-cal layer payload segment contains up to 45,408 data bits.The large packet size improves the estimation of time-varying channel responses in the received signal but intro-duces 20ms payload latency, resulting in a round trip delayof 70ms. It is desirable to explore ways to balance theneed for latency versus channel robustness and throughputin determining the physical layer packet sizes and algo-rithms.

Data Link Layer: No-Frills T-WWiN Bridge

In the first phase, the SDR is statically configured to sup-port the equivalent of a dedicated point to point link be-tween two network segments (e.g. ship to ship). The pairof SDRs acts as a single transparent bridge to route layer 2traffic between the networks they access. In this simplecase, the bridge forwards every layer two frame to its peerdevice. There is no MAC layer learning or filtering; allframes are simply forwarded. The SDR network interfacedriver is responsible for parsing the physical layer datastream and unpacking the encapsulated 802.3 frames. To

't- improve robustness in delimiting each Ethernet frame, theta data link layer adds a start of message (SOM) sequence,1 length and checksum to the beginning of each Ethernet[P frame as shown in Figure 8. Though not shown in the fig-al ure, the DIX (Xerox) Ethernet frame format is also sup-st ported.er.d

F-

1-

iste

T-WWiN Encapsulation of IEEE 802.3 Ethernet Frames

Figure 8 - Ethernet Frame Encapsulation

(C 2005 General Dynamics. All rights reserved.

-

4 of 7

Data Link Layer: Multi-Hop Switching

Much of the need for Ethernet network formation is inher-ently provided by improving the SDR layer 2 switchingcapability to include MAC learning, filtering and VLANsegregation. Learning and filtering corrects the problem ofunnecessary flooding of Ethernet communication packets.Once the SDR platform learns by observation, that a par-ticular MAC destination is reachable through a particularport, it no longer broadcasts frames with that destinationMAC, but rather directs them to the single correct interface(RF or Ethernet). This is beneficial even when the SDRchannel is configured as a dedicated bridge between twosegments of the same LAN.

In order to prevent network packet broadcast storms, alllayer 2 loops must be strictly avoided. This is typicallysolved by running the Spanning Tree Protocol (STP) on allswitches. This solution is non-ideal for at least two rea-sons; it introduces additional STP overhead onto the wire-less links, and it effectively disables all paths contributingto loops. The latter effect is an unacceptable loss of band-width in the wireless point to point network.

Support for trunked VLANs across the SDR network hasseveral potential advantages. Separating (looped) connec-tions onto separate VLANs allows them to operate in par-allel without creating broadcast storms. Support for paral-lel links is crucial for scalable bandwidth between hightraffic nodes and may also be beneficial in segregatingtraffic based on QoS levels. Use of layer 2 networks isalso convenient in that the SDR has no routing configura-tion overhead and end nodes can be directly connected tothe switched network with no IP reconfiguration. Finally,the use of layer 2 switching in the SDR permits rapid re-laying of frames between SDRs without invoking layerthree protocols (which increases likelihood of extendingcommunication beyond the modem/rf modules to networkspecific modules, implying increased latency within theSDR). On the down side, the efficient and accurate con-figuration of VLANS amongst the SDRs and external de-vices is an issue to be reckoned with.

The test bed provides an environment for comparing theperformance of traditional STP protected switched LANs,carefully segregated trunked VLANs and custom loopavoidance algorithms. Note that IP multicast (IGMP) sup-port can be provided by layer 2 network packet forwardingfor free as long as IP multicast MAC addresses are not fil-tered by the SDR RF network. Additionally, interestedmulticast peer devices must support the multicast packetfilter function at the layer 2 Ethernet MAC layer.

Network Layer: Multi-Hop Routing

The use of layer 3 routing can be used in place of, or inaddition to the layer 2 switching described in the previoussection. Furthermore, the routing function can be internalor external to the SDR. Each channel of the SDR is aroutable network interface over which the router will dis-cover other routers, assess link cost and share routing in-formation with other routers. Unlike switched networks,the routed topology is not susceptible to broadcast stormsbecause each radio channel represents a separate broadcastdomain.

Each SDR link can be characterized for reliability, signalstrength, etc. This information can be provided as a costmetric for internal/external routers to build optimal routesacross subsets of the SDR network. The phrase "SDRnetwork" refers to the collection of multi-hop inter-connected SDRs forming their own network cloud. TheSDR network is free to choose its own preferred routingprotocol without jeopardizing its interoperability with therest of the network domain. For example, the SDR net-work may opt to use RIP while the external network usesOSPF. Various mobile ad-hoc network (MANET) proto-cols will be evaluated on the test bed. In the layer 2switched network configuration, once a packet enters theSDR network, its route is controlled by the learned VLANpath not the edge router.

Initially, all test bed development will focus on the use ofexternal routers and switches to gain maximal capabilitywith minimal internal development. The longer term QoSbenefits will be gained by tightly integrating the switchingand routing functions with the radio wireless operation andperformance.

INTEGRATED QOS

Achieving sufficient QoS in the network is essential formeeting user expectations and for maximizing utilizationof network and RF bandwidth. This section introduceshow QoS objectives are supported through the SDR net-work at layers 1-3 of the OSI model.

Network Layer QoS

As noted in the background section, the major future mili-tary network architectures rely on HAIPE encapsulationand encryption across black networks, OSPF version 3routing and IP DiffServ Code Point (DSCP) based QoS.Differentiated Services (DiffServ), described in [8], is a setof technologies that allow network service providers tooffer services with different kinds of network QoS objec-tives [9]. The premise of DiffServ networks is that routerswithin the core of the network handle packets in differenttraffic streams by forwarding them using different per-hopbehaviors (PUBs). A PUB is defined as the queuing disci-

(C 2005 General Dynamics. All rights reserved. 5 of 7

pline and parameters associated with a particular queuingdiscipline. The PHB to be applied is indicated by theDSCP marking in the IP header of each packet. The DSCPmarkings can be applied by a trusted upstream node or bythe edge routers on entry to the DiffServ network.

The test bed is focused on implementing QoS behaviors inthe SDR that are compatible with the overall system QoSobjectives and mechanisms. It does not attempt to proposeor implement solutions to the system level QoS issues ofclassification, marking, conditioning and admission con-trol.

Data Link Layer QoS

Enforcement of a QoS service level agreement in a SDRwireless network is difficult to guarantee. Issues such astransmission quality and mobility that result in networktopology changes effect available network resources. Al-though enforcement of a QoS level may be difficult toguarantee, significant improvements can be realized byimplementation of layer 1 and 2 QoS in the SDR.

To effectively implement QoS across radio links betweensegments of the black private network, the DSCP QoSsupported by each segment must be complemented withequivalent QoS measures in the SDR lower layer proto-cols. In the SDR test bed, this is achieved by mapping thelayer 3 DSCP and layer 2 802.11P QoS fields into appropri-ate layer 1 and 2 SDR QoS behaviors. For example, pack-ets routed through a particular SDR interface extend thelayer 3 DSCP priorities into layer 2, 802.11P priorities,which are further extended into layer 1 priority queuesmapped directly to the 802.1P priority levels.

802.1P/Q as described in [10] is a technology that allowsnetwork switch infrastructures to prioritize & segmentnetwork traffic. 802. 1P/Q allows new RF & networkequipment to more easily integrate into existing (or new)network infrastructures. This technology is supported onBOTH IPv4 & IPv6 networks and provides a method bywhich RF bandwidth can be associated to network QoSweighting that occurs beyond the view of a router and atthe same time, allows a "canonical" interface into existing& new network infrastructures.

In addition to implementing layer 1 and 2 QoS in the SDR,the SDR can feed link quality metrics (e.g., max data rate,current data rate, latency, relative link quality) up to therouter (internal or external to the SDR). This significantlyenhances the routers reaction time in response to changesin transmission quality.

Since the Ethernet interface on an SDR is much faster thanthe RF links it supports, queuing to the Ethernet interfacesis not required. For Ethernet to RF bridging of data pack-ets, the SDR provides queuing of packets to the RF. In a

weighted round robin (WRR) scheme, each queue repre-sents a different packet classification group. The allowedqueue depth is configurable. The prioritization is definedby an integer number that defines the number of packetstaken off each queue in a round robin fashion. Packets thatoverflow the queue are dropped. This is an example of thetypes of issues that must be addressed when integratingrelatively slow wireless connections into a much fasterinfrastructure network infrastructure.

Physical Layer QoS

The physical layer implementation can be adapted to opti-mize particular goals. As noted earlier, the current T-WWIN payload processing introduces a minimum of 20mslatency per RF endpoint. Smaller T-WWIN packets woulddramatically lower this latency in exchange for signal re-covery quality and throughput. A detailed discussion ofthe physical layer optimizations are outside the scope ofthis paper. However, it is anticipated that the SDR testbed, already shown to be highly adaptable, will providefertile ground for exploring adaptive physical layer QoStechniques. The concepts highlighted below are suggestedfor future consideration.

Link establishment at optimum link speed - Deriving op-timum bidirectional link speed requires both sides of apeer RF link to test RF packet transmissions. Specifically,this requires sending a set of test packets between eachassociated peer to determine optimum modulation & RFlink speed attributes. Success or failure determineswhether RF link establishment is changed to a differentmodulation type (e.g. 16 QAM vs. 64 QAM).

RF frequency topology - Characterize the frequency band-width utilized by other users in the same band and dy-namically negotiate a frequency that is optimum for theuse of OFDM and avoid parts of the frequency band thatare being used. This improves QoS by improving the like-lihood of successful RF transmissions.

LOS NETWORK TOPOLOGY MANAGEMENT

To explore network management issues associated withintegrating network segments across low mobility WAN-like radio links, the test bed incorporates an automatednetwork topology management tool, called the DistributedNetworking Agent (DNA), running on a Linux laptop.DNA provides capabilities for planning and controllingphysical radio links based on current network configura-tion and performance of the communication nodes.

In this test bed DNA uses an 802.11 wireless network tosimulate a low bandwidth (<16k) control plane (i.e., a widearea coverage broadcast transmission media). Alterna-tively, the control channel can be provided through the useof one SATCOM DAMA timeslot. The DNA discovers

(C 2005 General Dynamics. All rights reserved. 6 of 7

the configured SDR radio assets and disseminates this andother configuration information to peer DNAs via the con-trol plane. DNA negotiates the preplanned network con-nectivity and issues the radio configuration commandsnecessary to create the needed links. Once initial discoveryand initial network formation is complete, the DNA com-municates with other DNA agents via the establishedcommunication links, referred to as in-band control, ratherthan using the low-bandwidth control plane. Traffic isinjected using an available traffic generator. To exercisethe link and topology management capabilities, an existingRF link is degraded to a point where it cannot effectivelypass traffic (e.g., high BER). DNA, which is continuallymonitoring the radio link quality parameters, automaticallydetects the degradation/failure and negotiates to create anadditional RF link to the destination. The router is in-formed by the radio, through the PPPoE protocol, of thenew better quality link and reroutes the traffic to that pathusing its own internal routing protocols.

1 Joint Vision 2020, Published by the U.S. Government Print-ing Office June 2000

2 Defense Acquisition Guidebook Chapter 7 Global Inforna-tion Grid, December 2004.

3 COL Jonathan A. Maddux and Dr. Gerardo J. Melendez,"The Network The Key to Transformation", Army AL&TMagazine January - February 2004, p12.

4 Warfighter Information Network-Tactical(WIN-T) Conceptof Operations [DRAFT] US ARMY SIGNAL CENTER 19November 1999

5 R Casey, "QoS Architecture for Navy Battle Groups", NavyVirtual Program Office web portal., September 2001.

6 J. Kleider, S. Gifford, K. Nolan, Derrick Hughes, and S.Chuprun, "Demonstrating Robust High Data Rate Capabil-ity on Software Defined Radio Using Anti-jam WidebandOFDM Waveforns," inproc. ofMILCOM, Oct. 2005.

7 D Cohlman, G. Osborn, "Feasibility and Roadmap for SCA,Wideband, and Networking Technology Insertion Into AFielded SDR", in proc ofMILCOM, Oct. 2005.

8 Y. Bernet, S. Blake, D. Grossman, A. Smith, "An InformalManagement Model for Diffserv Routers", RFC 3290, May2002.

9 Blake et al, "Architecture for Differentiated Services", IETFRFC 2475, December 1998.

10 IEEE 802. lp "Standard for Local and Metropolitan AreaNetworks - Supplement to Media Access Control (MAC).

SDR Wireless Test Bed Routes IP Based Voice and Data

Figure 9 - Network Management Test Bed Setup

SUMMARY

In summary, the DOD's transformational vision is one ofseamless real-time and near real-time global connectivitythrough cross echelon communication capabilities. Transi-tional test beds are needed to evaluate networking objec-tives against simulated results. This test bed will identifyprotocol related issues and further our understanding of theproposed notional architectures by verification throughimplementation. While simulation of larger networks is avaluable approach, the test bed is intended to provide atransition platform and network test bed on which toevaluate a subset of the protocols and techniques, with anemphasis on QoS, as required by the armed forces in theproposed future architectures. This test bed is particularlyfocused on demonstrating the interoperability of commer-cial devices and fielded SDR technology to achieve IPQoS objectives across small (<25 nodes) wireless radionetworks.

(C 2005 General Dynamics. All rights reserved. 7 of 7