Professional Documents
Culture Documents
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 139
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 140
Mobility patterns: Some tactical aircraft communications. Another key difference from
The physical layer is maintain predictable orbits to form communica- terrestrial networks is the much higher relative
tions relays while others fly attack runs. When velocity between nodes, potentially causing sig-
responsible for vari-
aircraft are engaged, the mobility patterns can nificant Doppler frequency offsets for which the
ous encoding and be very sporadic. While many studies have been receiver must compensate.
signaling functions performed on MANET routing protocols with The large region over which an aircraft has
random waypoint mobility and trajectory-based LOS coverage allows the communication links to
that enable transmis- mobility, much less work has been done in pre- span large distances. Design of airborne commu-
sion and reception of dictable orbit mobilities with realistic aircraft nications systems must balance the tradeoff of
maneuvering. using omni-directional communications versus
bits over a wireless
Spectrum considerations: Spectrum availabil- directional communications techniques. Omni-
medium. Two of the ity is limited, so new systems must coexist with directional systems simplify connectivity by low-
major challenges in existing systems. Efficient use and sharing of ering coordination overhead, but require high
available spectrum is highly desired. power transmissions to close long-distance links.
designing physical Operation in adversarial environments: Air- Additionally, omni-directional systems cause sig-
layers for next-gener- borne tactical networks operate in environments nificantly higher interference to neighbors.
where adversaries might actively target the sys- Directional systems can provide additional gain
ation ATNs are the
tem. This may require employment of interfer- and minimize interference, however, antenna
airborne wireless ence mitigation, jam resistance, and secure size constraints may require the use of higher
channel and interfer- protocols. carrier frequencies which suffer more from the
High cost of platform integration: Installing Doppler shift. Additionally, the use of direction-
ence resiliency. new RF hardware and antennas on existing air- al antennas requires accurate pointing, which
craft can incur significant integration costs. Evo- may be complicated by the high mobility of the
lution of new system technologies may be aircraft.
facilitated by reusing existing antennas and RF In ATNs the path loss may vary by over 50 dB
hardware. Due to the high cost of platform inte- due to the different ranges of neighboring air-
gration, the life of many existing systems are craft. Link adaptation may be needed to handle
decades long. Deployments of new designs must such a wide range of channel conditions. Using
have an integration path where new and old sys- higher code rates and more spectrally efficient
tems coexist. modulation when distances are small and path
The unique domain characteristics governing loss is relatively low can improve spectral effi-
ATNs help frame design constraints and consid- ciency. Spatial reuse increases (and hence over-
erations. In the following section, we will identify all network capacity improves) by reducing
how these domain characteristics impact the power levels when transmitting to nearby nodes
design of the first three layers of the network [3]. However, even with power and rate adapta-
stack. tion, the radio still needs to be able to discern
near and far users. Due to the extreme ranges in
signal-to-noise ratio (SNR), the radios will
DESIGN CONSIDERATIONS require a sophisticated analog front end.
Informed by the ATN domain characteristics, The LOS range of the aircraft may make
this section examines the key design challenges, them susceptible to interference from other
potential solutions, and areas of future research ground or airborne systems operating in the
at the physical layer, the link/MAC layer, and same band, and the airborne network may sim-
the network layer. ilarly cause interference to these other systems
(Fig. 2). As an example, Link 16 operates in
PHYSICAL LAYER DESIGN CONSIDERATIONS the aeronautical radio navigation bands
The physical layer is responsible for various (960–1215 MHz) for ground/aircraft and satel-
encoding and signaling functions that enable lite and employs techniques to minimize inter-
transmission and reception of bits over a wire- ference to these systems [4]. Attempts to reduce
less medium. Two of the major challenges in the potential for interference by limiting use of
designing physical layers for next-generation particular channels to certain geographic regions
ATNs are: must account for both the large LOS distances
• The airborne wireless channel. and the high mobility of the aircraft. A further
• Interference resiliency. complication for tactical airborne networks is
In this subsection, we detail these two issues and that the interference may be hostile and not
how they affect the design of ATN physical lay- amenable to mitigation through coordination.
ers. Interference resiliency can be achieved through
The channel in an airborne MANET differs waveform design (e.g., spread spectrum tech-
significantly from that in a terrestrial network. In niques [5]). Multiuser detection (MUD) can also
ground networks, the channel between two nodes be beneficial if interference is due to other net-
tends to be dominated by fading. In ATNs, how- work nodes. Successful use of MUD brings many
ever, two communicating aircraft will often have challenges, such as the complexity of processing
line-of-sight (LOS) connectivity even if separat- large numbers of users and, if ad hoc networking
ed by several hundred kilometers. Although is used, the potential lack of coordination
there may be limited multipath and fading due between nodes [6]. Although MUD is often con-
to ground reflection [1] and aircraft antenna sidered in conjunction with direct-sequence code
blockage [2], which are highly dependent on air- division multiple access (CDMA), it can also be
craft geometry, antenna placement, and ground applied to other signaling techniques such as fre-
terrain, these effects are generally much lower quency-hop spread spectrum. In this case, the
than those experienced in ground-to-ground MUD receiver may need to handle only a small
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 141
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 142
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 143
Legacy systems
Future ATNs can also leverage MANET (custom) (custom) (custom)
technologies to enable automatic node discovery
and route self-healing. As a result, network flexi- Link Link Link
bility and range are increased significantly. In (custom) (custom) (custom)
order to make this modular, IP-converged design
approach feasible for ATNs, several key compo- Phy Phy Phy
nents are needed: (custom) (custom) (custom)
• Header compression.
• Radio-to-router interface (R2RI) definition.
• Multi-topology MANET routing.
In tactical networks, many of the messages
are short. By moving to IP, headers may domi-
nate the traffic and waste precious resources. A
voice-over-IP (VoIP) transmission between two Router
aircraft might incur 60 bytes of header (40 bytes
IPv6, 8 bytes UDP, 12 bytes RTP) for a 10-18 Network layer
(modular)
byte message, depending on the codec used.
Next-generation ATNs
Employing commercial header compression
techniques can potentially reduce this to 4-6 Pass cross
Common
bytes. Additionally, pushing IP packets through interface layer info
Link 16 is inefficient because IP headers con-
sume much of the limited Link 16 capacity [13].
In previous work [14], commercial IP header Link Link Link
compression protocols as well as a new proposed PHY PHY PHY
stateless header compression approach, MANET
IP Header Compression (MIPHC), were evalu- Radio/ Radio/ Radio/
ated for suitability in the MANET environment. waveform waveform waveform
It was shown that commercial header compres-
sion techniques work fairly well in a MANET Figure 4. IP convergence, a common path to passing cross-layer informa-
environment because of the ability to re-use pre- tion, and routing with heterogeneous link technologies are key aspects of
viously built state between compressor and the network layer design.
decompressor even though the topology is regu-
larly changing. However, these techniques suffer
from potential security vulnerabilities on state information to optimize and tailor routing deci-
setup and refresh. MIPHC did not provide as sions to traffic flows. Similarly, higher-layer
much compression gains as the commercial information provided to the link layer allows the
stateful protocols, however it provided lower link layer to optimize link resource allocation
latency operation and resilience toward attack. based on traffic priority or end-to-end delivery
Cross-layer optimization can be achieved requirements.
through two main methods, as shown in Fig. 4: To enable the cross-layer optimization with
• Tightly coupling the network stack and physically separate and diverse radio and router
allowing multi-hop routing decisions to components, a standardized radio-to-router
access link layer information directly. interface (R2RI) is needed. Several R2RI tech-
• Modularizing key components by defining niques are explored in [16], and it was discov-
the radio or link layer to provide the best ered that although access to instantaneous link
one-RF hop possible, using an external information speeds up reaction to link outages,
router to perform multi-hop routing, and modifying routes based on small changes to link
defining a standard radio-to-router inter- metrics can yield significantly higher overhead
face (R2RI) to pass relevant information and instability. Open areas of research include
between the radio and router. the definition of clear and appropriate standard-
One of the advantages of a tightly coupled design ized metrics, design and optimization of routing
approach is the potential for cross-layer design protocols to effectively use the link metrics, and
and optimization to enable efficient utilization of flow control strategies between the router and
the limited wireless resources. It is well known radio. Abstracting radio systems to a set of links
that by jointly optimizing routing and medium with associated link metrics through the R2RI
access control, significant capacity gains can be enables research into multi-topology routing
achieved [15]. However, this approach limits the protocols that leverage a heterogeneous set of
system to homogeneous nodes and links. links and link metrics to provide efficient, stable,
With a modular network architecture, effi- and resilient routing over multiple radio paths.
cient and informed use of the multiple heteroge- Although there are some schemes that have
nous links can be enabled by passing cross-layer been proposed [17], a holistic test with represen-
information between the radio and the router. tative links is still needed.
Link quality, reliability, availability, SNR, and ATN nodes must be able to relay information
other statistics affect the usability of a link, and to enable longer range and connectivity. Due to
an informed network layer can leverage this their long transmission ranges (100s of kilome-
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 144
SUMMARY
Design Consideration Research Challenges
Table 1 highlights the key domain characteris-
–Techniques to share spectrum with long-range tics, design considerations, and research chal-
Spectral reuse (Physical omni-directional transmissions. lenges associated with each layer of the network
Layer) –Link adaptation (power and rate) to provide stack as it pertains to ATNs.
spectral efficiency.
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.
CHENG_LAYOUT_Layout 5/7/14 1:12 PM Page 145
[8] B. Suman, L. Mangal, and S. Sharma, “Analyzing Impact cal Engineering from Clemson University, Clemson, SC. His
of TDMA MAC Framing Structure on Network Through- research interests include spread-spectrum communica-
put for Tactical MANET Waveforms,” Conf. Advances in tions, multiuser techniques, and packet radio networks. Efficiently leveraging
Commun. and Control Systems (CAC2S), 2013.
[9] D. Bertsekas and R. Gallager, Data Networks (2nd Ed.), BENJAMIN “RUSS” HAMILTON (benjamin.hamilton@ll.mit.edu)
multiple heteroge-
Upper Saddle River, NJ, USA: Prentice-Hall, Inc., 1992. received his B.S. in Electrical Engineering from Auburn Uni- neous radio systems
[10] D. Ripplinger, A. Narula-Tam, and K. Szeto, “Schedul- versity in 2005, his M.S. and Ph.D. in Electrical and Com-
ing vs. Random Access in Frequency Hopped Airborne puter Engineering from the Georgia Institute of Technology is a significant chal-
Networks,” IEEE Military Commun. Conf., MILCOM in 2007 and 2012. While at Georgia Tech, he performed
2012, 2012, pp. 1–6. research on synchronization and channel estimation in lenge. By abstracting
[11] J. Hammond and H. Russell, “Properties of a Transmis- wireless systems and focused on distributed estimation in
sion Assignment Algorithm for Multiple-hop Packet wireless networks for his dissertation. After receiving his
radios to a set of
Radio Networks,” IEEE Trans. Wireless Commun., vol. 3, Ph.D., he joined the Communications division of MIT Lin- links and associated
no. 4, 2004, pp. 1048–52. coln Laboratory as a member of the Technical Staff. His
[12] A. Muqattash, M. Krunz, and W. E. Ryan, “Solving the research focuses primarily on communications and net- link metrics, efficient
Near-Far Problem in CDMA-based Ad Hoc Networks,” working in the airborne tactical domain.
Ad Hoc Networks, vol. 1, no. 4, Nov. 2003, pp. 435–53. multi-topology
[13] Y. Yoon et al., “Header Compression for Resource and DAVID RIPPLINGER (david.ripplinger@ll.mit.edu) is an Associ-
Energy Efficient IP over Tactical Data Link,” Springer ate Staff member in the Airborne Networks Group at MIT
routing mechanisms
Future Generation Information Technology (FGIT) 2010, Lincoln Laboratory. His research interests include theoreti- can be employed
vol. 6485, 2010, pp. 180–90. cal optimization, design, simulation, and testing of new
[14] B.-N. Cheng et al., “MANET IP Header Compression,” protocols for wireless networks, with an emphasis on to route over
IEEE Military Commun. Conf., MILCOM, 2013. medium access control (MAC) design. Recent work has
[15] M. Chiang et al., “Layering as Optimization Decompo- focused on characterization of the behaviors of random heterogeneous links
sition: A Mathematical Theory of Network Architec- access and scheduling MACs in a frequency hopping envi-
tures,” Proc. IEEE, vol. 95, no. 1, pp. 255—312, Jan. ronment, as well as the design of low-overhead, distribut-
simultaneously.
2007. ed random access protocols. He received an M.S. in
[16] B.-N. Cheng et al., “Comparing Radio-to-Router Inter- Computer Science and a B.S. in Physics and Spanish Trans-
face Implementations on Experimental COTS and Open lation from Brigham Young University.
Source Routers,” IEEE Military Commun. Conf., MILCOM
2012, Oct. 2012. CHAYIL TIMMERMAN (ctimmerman@ll.mit.edu) is a member of
[17] C. Dearlove and T. Clausen, “Multi-Topology Extension Technical Staff in the Advanced Satcom Systems and Oper-
for the Optimized Link State Routing Protocol version 2 ations Group at MIT Lincoln Laboratory. His research inter-
(OLSRv2),” IETF, Internet Draft (work in progress) 01, ests include design, development, and implementation of
2013. advanced communications systems. Recent work has
[18] B.-N. Cheng and S. Moore, “An Evaluation of MANET focused on tactical communications systems, high band-
Routing Protocols on Airborne Tactical Networks,” IEEE width satcom systems, and dynamic code rate adaptation.
Military Commun. Conf., MILCOM 2012, Oct. 2012. He holds a M.S. from Villanova University and a B.S.E.E.
from Drexel University.
BIOGRAPHIES LEONID VEYTSER (veytser@ll.mit.edu) is a member of the Air-
BOW-NAN CHENG (bcheng@ll.mit.edu) is a member of Tech- borne Networks Group at MIT Lincoln Laboratory. His
nical Staff in the Airborne Networks Group at MIT Lincoln research interests include routing and distributed comput-
Laboratory. His research interests include design, develop- ing in disadvantaged wireless networks. His recent work
ment, prototyping, and test and evaluation of next genera- has focused on enhancing application performance at the
tion routing and information disseminations solutions for tactical edge as well as radio-aware routing in airborne
airborne backbone and tactical networks. Recent work has and tactical networks. He received his B.A. and M.A.
focused heavily on radio-aware routing, which leverages degrees in Computer Science at Boston University.
link layer information at the network layer to enhance
multi-hop MANET routing. He received M.S. and Ph.D. A RADHANA N ARULA -T AM (arad@ll.mit.edu) is an Assistant
degrees in computer systems engineering from Rensselaer Leader of the Airborne Networks Group at MIT Lincoln Lab-
Polytechnic Institute and holds a B.S. degree in electrical oratory. Her research interests include topology design,
engineering from the University of Illinois at Urbana-Cham- resource scheduling, MANET capacity, networking proto-
paign. cols, and Quality of Service. She has worked on a variety of
systems including optical networks, satellite communica-
FREDERICK J. BLOCK (fblock@ll.mit.edu) is a member of the tion systems, and tactical wireless communication systems.
Technical Staff in the Advanced Satcom Systems and Oper- She holds a B.S.E. degree from the University of Pennsylva-
ations Group at MIT Lincoln Laboratory. He received the nia and S.M. and Ph.D. degrees from the Massachusetts
B.S. (summa cum laude), M.S., and Ph.D. degrees in Electri- Institute of Technology, all in electrical engineering.
Authorized licensed use limited to: University of Defence. Downloaded on December 14,2022 at 21:23:44 UTC from IEEE Xplore. Restrictions apply.