Secrets in The Sky: On Privacy and Infrastructure Security in DVB-S Satellite Broadband
Secrets in The Sky: On Privacy and Infrastructure Security in DVB-S Satellite Broadband
277
WiSec ’19, May 15–17, 2019, Miami, FL, USA James Pavur, Daniel Moser, Vincent Lenders, and Ivan Martinovic
observed that the satellite transmitted DVB-S encapsulated Inmarsat’s BGAN system), this makes DVB-S a particularly
web-browsing data in the clear. The researchers noted that attractive target for attackers [20].
these short-comings may have been the responsibility of end- The DVB-S standard transmits data per the Moving Pic-
users who relied on unencrypted protocols such as POP3 tures Experts Group (MPEG) standards in the form of
emails [5]. MPEG transport streams (MPEG-TS) [21]. While MPEG-
A few years later, at Black Hat 2009, private security TS is primarily used for media broadcasting, the standard
researcher Adam Laurie presented on a traffic interception has also been extended to support many other types of data.
experiment which used modified equipment provided by a In particular, one such extension, called “Digital Storage
satellite ISP [6]. This was followed at Black Hat in 2010 Media Command and Control” (DSM-CC) was developed
by Leonardo Nve Egea, who demonstrated satellite internet to provide interactive features on Video Cassette Recorders
sniffing using DVB PCI cards in the Ka-Band [7]. Some pe- (VCR) [22]. As demands for interactive satellite broadband
ripherally related academic work has emerged in the form services grew, DSM-CC was repurposed to relay arbitrary
of various standards revisions for satellite internet [8, 9, 10]. packetized data to and from satellite internet customers via an
Moreover, some related non-academic work on other aspects additional encapsulation layer called Multiprotocol Encapsu-
of satellite systems (such as the security of software on satel- lation (MPE) [23]. A subsequent revision of the MPE method
lite terminals) has taken place [11, 12]. However, the past called Unidirectional Lightweight Encapsulation (ULE) has
eight years have seen very little study of satellite broadband also been created which allows for the transmission of packe-
transmissions and more than a decade has elapsed since major tized data without the use of DSM-CC tables [23]. In both
academic consideration of these networks. protocols, IP packets destined for many distinct customers
Beyond academia and security conferences, hobbyist and are transmitted on the same stream and then extracted by
criminal communities are the primary source of modern customer equipment on the basis of address information in the
domain-specific research. For example, online communities IP header or ISP-assigned MAC addresses in the MPE/ULE
dedicated to the receipt of free satellite television have devel- headers.
oped various high-quality tools for scanning, intercepting, and
interpreting DVB-S signals [13, 14]. Online forums dedicated 3.1 Data Visibility
to the illegal cracking and cloning of private keys associated The specific topology of satellite networks can have significant
with satellite television networks are active hotbeds of infor- impacts on an attacker’s ability to understand broadband
mal offensive security research [15]. Finally, criminal groups transmissions. Our threat model takes into account two of
have demonstrated the usage of satellite internet connections the most common network topologies.
for exfiltrating data to undetectable command-and-control
stations [16].
In short, only a small body of high-quality but dated aca-
demic research exists. Researchers in the 2000s suggested
severe security shortcomings in DVB-S broadband but no as-
MPE Two Way Traffic
sessment has been performed to assess if the modern situation ISP Transponder
has improved or changed [5, 6, 7]. This paper updates and
expands previous findings in the context of modern internet
traffic. Unlike prior research, which focused on individual
Attack Attack
satellites and providers, we focus on a broad cross-section of Position
Position Internet
the industry and present what we believe to be the first large- Home User 2
1 Satellite Ground Station
scale multi-satellite security study of DVB-S broadband.
278
Secrets in the Sky WiSec ’19, May 15–17, 2019, Miami, FL, USA
downlink-to-ISP connection (Position 2 in Figure 2) would be specialized hardware targeted to their ISP is used in the
able to view requests made to the internet. Depending on the form of a satellite receiver/modem. Relying on generic equip-
specific geographic location of the attacker in either scenario, ment can increase processing errors – especially for complex
additional traffic may also be visible through interception of modulation modes such as 16 and 32-APSK.
signals emanating from antenna side lobes.
The second configuration offered by some specialized ISPs
such as Broadsat’s Opensky combines satellite and terrestrial 4.2 Deployment
linkages (Figure 3) [24]. Uplink requests are transmitted via Two inexpensive satellite receiver assemblies (of the sort in-
a terrestrial connection, which typically has better latency tended for RV-camping) were deployed to simulate this threat
(Position 1 in Figure 3). Requests are sent to a proxy operated model. One consisted of a 75 cm, flat-panel satellite receiver
by the satellite ISP (Position 2 in Figure 3) and responses re- dish and a TBS-6983 DVB-S receiver. The other consisted of
layed back to the user via satellite (Positions 3-5 in Figure 3). a 60 cm flat-panel dish, a motorized targeting assembly, and
Such configurations are ideal in cases where uplink latency is a TBS-6903 DVB-S receiver. The 75 cm dish remained in a
of higher priority than uplink bandwidth and for customers fixed position while the 60 cm dish was repositioned to target
with extant but inadequate terrestrial service. Here, an at- many satellites over the duration of the study. The panels
tacker would be incapable of observing uplink traffic over the were configured to receive Ku-band transmissions between
air as it is transmitted via wire rather than radio. 10,700 MHz and 12,750 MHz with both vertical and horizon-
tal polarizations. Both assemblies were located in Europe and,
due to environmental constraints, could observe geostationary
satellites positioned between 40°East and 37°West.
A set of 14 geostationary satellites were selected based on
Downlink ISP Transponder signal quality at the collection site. From these satellites over
5
4 350 transponders were identified using existing “Blind Scan”
tools.
A collection of Python utilities developed for the purpose
DSL/Dial-Up Uplink 3 of this study was used to analyze each of these transponders
Home User 1 ISP Proxy for signs of DVB-based internet transmissions on the basis of
Satellite Ground Station
2 three criteria. First, a stream was deemed more likely to carry
internet traffic if DSM-CC (MPE) services were listed in the
Internet stream’s program table. Second, streams which contained
valid UDP or TCP packets, based on existing MPEG-TS
dissectors in Wireshark, were flagged as candidates. Finally,
Figure 3: A combined terrestrial uplink, space downlink satel- streams were parsed against a list of regular expressions
lite internet setup commonly seen in internet traffic. A total of 19 streams met
at least two of these criteria and 4 additional matches were
identified on the basis of the regular expression engine alone.
4 EXPERIMENTAL DESIGN From these 23 transponders, streams which appeared to carry
4.1 Attacker Equipment IP-TV traffic or simple device firmware update services were
discarded, along with streams with extremely low signal
For this study, we sought to discern the capabilities of a
quality and throughput (anything less than 5 kb/s). This left
single malicious individual rather than a larger organization
13 transponders for further study.
or nation state. As such, we restricted our equipment selection
to hardware that was readily available for purchase online and
only employed free software tools. The total cost of necessary 5 DATA COLLECTION
equipment was under e300, as demonstrated by the budget
In June, 2018, approximately five hours of traffic were recorded
in Table 1.
on each of the 13 selected transponders. These recordings
Table 1: Hypothetical Attacker Budget were initiated automatically in sequence over the course of
three days. Each recording was processed to remove NULL
packets and irrelevant program data (such as programme
Equipment Cost data related to satellite television). After this process we were
Selfsat H30D Satellite Dish e85 left with 50 gigabytes of satellite internet traffic. Depending
TBS 6983 Satellite PCI-E Card e197 on location, radio conditions, and transmission modes, the
3-Meter Coaxial Cable e3 amount of data collected per transponder ranged from as low
Total e285 as 8 megabytes to as high as 10 gigabytes (see Figure 4).
As anticipated in Section 4.1, recordings included many
It is worth noting that equipment quality can have a data errors resulting from our use of general-purpose hobby-
meaningful impact on capabilities. For legitimate customers, ist equipment. Nevertheless, sufficient information could be
3
279
WiSec ’19, May 15–17, 2019, Miami, FL, USA James Pavur, Daniel Moser, Vincent Lenders, and Ivan Martinovic
1000
6.1 Privacy
A surprising amount of sensitive information appeared in the
100
collected data. Indeed, many of the same categories identified
over a decade ago still appear in modern satellite traffic.
One significant improvement since the mid-2000s has been
increased adoption of SSL/TLS encryption. While this pro-
10 tects against certain types of eavesdropping attacks, the very
process of requesting and exchanging SSL certificates leaks
potentially revealing information. Our data included over
1 52,000 SSL “wildcard” certificates from around 1,200 distinct
1 2 3 4 5 6 7 8 9 10 11 12 13
domains (Figure 5). Information a user might consider deeply
Satellite ID private – such as TLS certificates or DNS responses from
various adult websites – is, in fact, being broadcast across an
Figure 4: Volume of internet traffic on each targeted transpon- entire continent. With collating data, such as knowledge of a
der during the five-hour analysis window. user’s IP address, this risk becomes particularly severe.
*.google.com
extracted despite this corruption to give a general characteri- *.gvt1.com
*.googleapis.com
zation of security concerns. *.g.doubleclick.net
*.adnxs.com
5.0.1 Ethics, Data Privacy, and Legal Considerations. Prior *.icloud.com
to our experiment, it was unclear what sort of information *.mail.me.com
[PRIVATE - ENERGY PROVIDER]
would be uncovered. As such, we assumed a worst-case sce- *.google-analytics.com
nario and treated all recorded radio signals as if they might *.dropbox.com
*.c.docs.google.com
contain sensitive information. Data was stored at the collec- *.criteo.com
tion site in Europe and both physical and electronic access *.crashlytics.com
*.1.oca.nflxvideo.net
was restricted. Local laws relating to the interception and *.googleusercontent.com
analysis of radio traffic were strictly adhered to. We also *.smoot.apple.com
*.pipe.aria.microsoft.com
made plans to responsibly disclose any security issues which *.twimg.com
warranted it to the appropriate authorities. After the study *.rubiconproject.com
all collected data was deleted. *.doubleclick.net
0 1000 2000 3000 4000 5000 6000
6 FINDINGS
Across all thirteen frequencies included in the final phase Figure 5: The top 20 domains identified on a basis of SSL
of our study, broadband traffic was transmitted in plain- certificates. Number 8 has been hidden as it is a private sub-
text. Of course, well-encrypted transmissions would not have domain range for a major energy provider.
been distinguishable from non-internet traffic and it is thus
unclear to what extent these thirteen service providers are
While SSL usage is widespread, our naive string-matching
representative of the industry as a whole. Nevertheless, having
analysis nevertheless uncovered thousands of unencrypted
the same issues appear over many distinct providers suggests
HTTP requests, file downloads, FTP sessions, torrent connec-
that the earlier single-provider studies discussed in Section 2
tions, VoIP conversations and emails. The chart in Table 2
were not merely anecdotal and the problems they identified
indicates which of these general classes of sensitive informa-
have not yet been addressed.
tion were identified on each of the 13 transponders. Due to
The dangers of unencrypted wireless transmissions are well
the sensitive nature of our findings, specific service providers
understood and, to some degree, academically uninteresting.
and satellite names have been withheld. These findings raise
However, unique properties of satellite broadband act as
legal and business concerns regarding whether it is the re-
novel risk-multipliers.
sponsibility of satellite service providers to protect customers
The principle differentiator is scale. Our experiment in-
using insecure protocols over DVB-S or if responsibility for
cluded data from a coverage footprint of more than 110
encryption in transit falls to end-users.
million square kilometers (Figure 1). A handful of strate-
gically located satellite dishes, would allow an attacker to 6.1.1 Demonstrative Cases. Beyond this broad sense of infor-
intercept broadband signals encompassing most of the globe. mation leakage, a number of individual “narratives” emerged
4
280
Secrets in the Sky WiSec ’19, May 15–17, 2019, Miami, FL, USA
281
WiSec ’19, May 15–17, 2019, Miami, FL, USA James Pavur, Daniel Moser, Vincent Lenders, and Ivan Martinovic
appeared in the streams, but an in-depth analysis of maritime from inspecting the necessary packet headers to continue
communications protocols was well beyond the project scope. providing these services.
In the early 2000s, several solutions to these issues emerged,
ranging from decrypting traffic at the satellite ISP ground
station and re-encrypting for transmission over the internet
7 POTENTIAL SOLUTIONS to implementing revisions to the IPSec protocol itself to ex-
Our findings suggest a need for improved communications tract TCP headers needed for PEP usage, however these have
privacy in satellite broadband. The current state of industry gained little traction [9, 26, 31]. While promising standards
has resulted in the leakage of sensitive information from both proposals exist for encrypted satellite data links in scien-
individual and industrial satellite broadband customers. tific missions, additional work adapting these to multi-user
The application of terrestrial encryption techniques to broadband ecosystems is needed [32].
satellite environments is non-trivial. Satellite transmissions Beyond work to bolster satellite internet encryption, future
cover vast distances and are subject to speed-of-light latency work characterizing these vulnerabilities may prove useful.
effects (upwards of 500ms for a round-trip transmission to While broader than previous research, our study still only
GEO) and packet loss which can impair the function of studied a handful of operators in the Ku-band and was re-
encryption schemes designed for high-reliability terrestrial stricted to the DVB-S protocol. Moreover, our experimen-
environments (e.g. by requiring re-transmission of corrupted tation was limited to passive eavesdropping attacks only.
key materials) [25, 26]. Moreover, satellites themselves are Research on the capabilities of an attacker to compromise
limited in terms of computing capabilities and any on-board not just the confidentiality but also the integrity of satellite
cryptographic operation risks trading off with other mission broadcasts may uncover additional status quo risks. For ex-
functionality [26]. ample, attacks against the layer-two routing protocols used in
In television networks, “scrambling” algorithms which ap- satellite networks may facilitate sophisticated session spoofing
ply encryption to an entire MPEG-TS programme are widely and hijacking attacks.
used to prevent piracy of premium television channels. How-
ever, these techniques are, at present, not well suited to 8 CONCLUSION
internet traffic for two reasons. First, many of the dominant Our experimental analysis raises significant concerns for the
algorithms in this space (e.g. the Common Scrambling Algo- security of DVB-S broadband. Severe confidentiality short-
rithm or PowerVu) have been demonstrated to have severe comings were identified across more than a dozen service
cryptographic weaknesses [27, 28, 29]. While these may be operators and several gigabytes of potentially sensitive web
acceptable for low sensitivity TV broadcasts, where the prin- traffic were collected in a matter of hours. An attacker with
ciple goal is often to simply increase the complexity of piracy, cheaply available hobbyist equipment may compromise the
internet traffic and sensitive data may merit more robust security and privacy of individuals in an area encompassing
protections. Moreover, because these scrambling algorithms hundreds of millions of square kilometers. Moreover, satel-
apply at the level of entire streams, customers necessarily lite eavesdropping provides a potential route to harming
share access to a “master key” (often updated at regular inter- many connected critical infrastructure systems such as power-
vals) which could be used to compromise the privacy of other generation facilities.
customers whose traffic is multiplexed into the same trans- We suggest that future cryptographic work which finds a
mission [29]. Future work which builds on these scrambling balance between satellite network performance and robust
techniques with a focus on the needs of internet customers encryption represents a critical step towards resolving these
may represent a possible avenue for developing more robust issues. This work might build off of progress in stream-level
DVB-S protections. scrambling techniques or in revisions to terrestrial encryption
One alternative approach would be the use of tunneling protocols that account for unique characteristics of satellite
mechanisms such as IPSec. In the short term, this is likely networks.
the best approach for individual customers and infrastruc- Secure satellite internet is vital to reaping the rewards of
ture operators. However, terrestrial tunneling technologies a connected future. In light of growth in the infrastructure
impose significant performance constraints over satellite con- and IoT space, the security community and satellite industry
nections [8]. Due to the significant latency in GEO broad- must cooperate to make satellite broadband private, reliable,
casting, satellite service operators have adopted connection and secure.
acceleration techniques which help minimize these effects.
Specifically, broadband providers widely make use of Per- REFERENCES
formance Enhancing Proxies (PEPs) to simulate artificial [1] H. D. Clausen, H. Linder, and B. Collini-Nocker. 1999.
TCP acknowledgments and TCP features which misinterpret Internet over direct broadcast satellites. IEEE Commu-
satellite latency as network congestion [26, 30]. Moreover, nications Magazine, 37, 6, (June 1999), 146–151. issn:
although TLS functions over satellite networks, plain-text 0163-6804. doi: 10.1109/35.769289.
HTTP requests will often appear more performant to users [2] Olivier Anstett. Internet Broadband: A New Source
due to the heavy application of HTTP accelerators [26]. Tun- of Growth. (2015). https://www.eutelsat.com/files/
neling and end-to-end encryption prevents satellite operators
6
282
Secrets in the Sky WiSec ’19, May 15–17, 2019, Miami, FL, USA
contributed / investors / pdf / Capital - Markets - Day - [14] crazycat69. 2018. CrazyScan: Satellite/terrestrial/cable
2015/Internet%20broadband_a_new%20_source_ scan software. Retrieved 06/20/2018 from https : / /
of_growth.pdf. sourceforge.net/projects/crazyscan/.
[3] David Grossman. 2018. The Race for Space-Based In- [15] World-Satellite. 2018. World-Satellite Forum. Retrieved
ternet Is On. (January 3, 2018). Retrieved 05/31/2018 06/20/2018 from http://www.world-satellite.net/.
from https://www.popularmechanics.com/technology/ [16] Stefan Tanase. 2015. Satellite Turla: APT Command
infrastructure/a14539476/the- race- for- space- based- and Control in the Sky. (September 9, 2015). Retrieved
internet-is-on/. 06/20/2018 from https://securelist.com/satellite-turla-
[4] Mohammad Marashi. 2017. Satellites are critical for apt-command-and-control-in-the-sky/72081/.
IoT sector to reach its full potential. (June 8, 2017). [17] ETSI. 2014. Digital Video Broadcasting (DVB); Sec-
Retrieved 05/31/2018 from http://social.techcrunch. ond generation framing structure, channel coding and
com/2017/06/08/satellites-are-critical-for-iot-sector- modulation systems for Broadcasting, Interactive Ser-
to-reach-its-full-potential/. vices, News Gathering and other broadband satellite
[5] André Adelsbach and Ulrich Greveler. 2005. Satellite applications. Retrieved 06/01/2018 from https://www.
Communication without Privacy - Attacker’s Paradise. dvb.org/standards/dvb-s2.
In Sicherheit. [18] Koen Williams. 2014. DVB-S2X Demystified. White
[6] Adam Laurie. $atellite Hacking for Fun & Pr0fit! (2009). Paper. Newtec, (March 2014). https://www.newtec.eu/
http : / / www . blackhat . com / presentations / bh - dc - frontend/files/userfiles/files/Whitepaper%20DVB_
09/Laurie/BlackHat-DC-09-Laurie-Satellite-Hacking. S2X.pdf.
pdf. [19] Comsys. [n. d.] VSAT Network Types. Retrieved 03/07/2019
[7] Leonardo Egea. Playing in a Satellite environment 1.2. from https://www.comsys.co.uk/wvr_nets.htm.
(2010). http://www.blackhat.com/presentations/bh- [20] Inmarsat. [n. d.] BGAN Voice and Broadband Service.
dc - 10 / Nve _ Leonardo / BlackHat - DC - 2010 - Nve - Retrieved 04/01/2019 from https://www.inmarsat.
Playing-with-SAT-1.2-wp.pdf. com/service/bgan/.
[8] S. Iyengar, H. Cruickshank, P. Pillai, G. Fairhurst, [21] International Telecommunication Union. 2014. H.222.0
and L. Duquerroy. 2007. Security requirements for IP Information technology – Generic coding of moving
over satellite DVB networks. In 2007 16th IST Mobile pictures and associated audio information: Systems.
and Wireless Communications Summit. 2007 16th IST (October 2014). https://www.itu.int/rec/T- REC-
Mobile and Wireless Communications Summit. (July H.222.0-201410-S/en.
2007), 1–6. doi: 10.1109/ISTMWC.2007.4299224. [22] Vahe Balabanian, Liam Casey, Nancy Greene, and
[9] Laurence Duquerroy, Sebastien Josset, O. Alphand, P. Chris Adams. 1996. An Introduction to DSM-CC. (No-
Berthou, and T. Gayraud. 2004. SatiPSec : An Op- vember 1996). Retrieved 06/01/2018 from http://www.
timized Solution for Securing Multicast and Unicast iuma.ulpgc.es/~nunez/procmultimedia98- 00/cselt/
Satellite Transmissions. In 22nd AIAA International mpeg/documents/dsmcc/dsmcc.htm.
Communications Satellite Systems Conference & Ex- [23] Teh Chee Hong, Wan Tat Chee, and R. Budiarto. 2005.
hibit 2004 (ICSSC). American Institute of Aeronautics A Comparison of IP Datagrams Transmission using
and Astronautics, (May 2004). doi: 10.2514/6.2004- MPE and ULE over Mpeg-2/DVB Networks. In 2005
3177. Retrieved 01/18/2019 from https://arc.aiaa.org/ 5th International Conference on Information Commu-
doi/abs/10.2514/6.2004-3177. nications Signal Processing. 2005 5th International Con-
[10] H. Cruickshank, M. P. Howarth, S. Iyengar, Zhili Sun, ference on Information Communications Signal Pro-
and L. Claverotte. 2005. Securing multicast in DVB- cessing, 1173–1177. doi: 10.1109/ICICS.2005.1689239.
RCS satellite systems. IEEE Wireless Communications, [24] Broadsat. [n. d.] OPENSKY - One way satellite inter-
12, 5, (October 2005), 38–45. issn: 1536-1284. doi: net. Retrieved 06/01/2018 from https://www.broadsat.
10.1109/MWC.2005.1522103. com/en/opensky/.
[11] Ruben Santamarta. 2014. SATCOM Terminals: Hack- [25] Dan Lester and Harley Thronson. 2011. Human space
ing by Air, Sea, and Land. DEFCOM White Paper, 26. exploration and human spaceflight: Latency and the
https://www.blackhat.com/docs/us-14/materials/us- cognitive scale of the universe. Space Policy, 27, 2,
14 - Santamarta - SATCOM - Terminals - Hacking - By - (May 1, 2011), 89–93. issn: 0265-9646. doi: 10.1016/
Air-Sea-And-Land-WP.pdf. j.spacepol.2011.02.002. Retrieved 06/15/2018 from
[12] Ruben Santamarta. 2018. Last Call for SATCOM Secu- http://www.sciencedirect.com/science/article/pii/
rity. Blackhat Whitepaper 2018, (August 2018). https: S0265964611000348.
/ / i . blackhat . com / us - 18 / Thu - August - 9 / us - 18 - [26] A. Roy-Chowdhury, J. S. Baras, M. Hadjitheodosiou,
Santamarta-Last-Call-For-Satcom-Security-wp.pdf. and S. Papademetriou. 2005. Security issues in hybrid
[13] Cjcr-Software. 2016. EBSpro. Retrieved 06/20/2018 networks with a satellite component. IEEE Wireless
from http://ebspro.net/. Communications, 12, 6, (December 2005), 50–61. issn:
1536-1284. doi: 10.1109/MWC.2005.1561945.
283
WiSec ’19, May 15–17, 2019, Miami, FL, USA James Pavur, Daniel Moser, Vincent Lenders, and Ivan Martinovic
[27] W. Li and D. Gu. 2007. Security Analysis of DVB [30] A. J. H. Fidler, G. Hernandez, M. Lalovic, T. Pell,
Common Scrambling Algorithm. In The First Interna- and I. G. Rose. 2002. Satellite — A New Opportunity
tional Symposium on Data, Privacy, and E-Commerce for Broadband Applications. BT Technology Journal,
(ISDPE 2007). The First International Symposium on 20, 1, (January 1, 2002), 29–37. issn: 1573-1995. doi:
Data, Privacy, and E-Commerce (ISDPE 2007). (No- 10.1023/A:1014561823985. https://doi.org/10.1023/A:
vember 2007), 271–273. doi: 10.1109/ISDPE.2007.63. 1014561823985.
[28] Erik Tews, Julian Wälde, and Michael Weiner. 2011. [31] M. P. Howarth, S. Iyengar, Z. Sun, and H. Cruick-
Breaking DVB-CSA. In Research in Cryptology (Lec- shank. 2004. Dynamics of key management in secure
ture Notes in Computer Science). Western European satellite multicast. IEEE Journal on Selected Areas in
Workshop on Research in Cryptology. Springer, Berlin, Communications, 22, 2, (February 2004), 308–319. issn:
Heidelberg, (July 20, 2011), 45–61. isbn: 978-3-642- 0733-8716. doi: 10.1109/JSAC.2003.819978.
34158-8 978-3-642-34159-5. doi: 10.1007/978-3-642- [32] CCSDS. 2018. Space Data Link Security Protocol -
34159-5_4. Retrieved 06/15/2018 from https://link. Summary of Concept and Rationale. Green Book. (June
springer.com/chapter/10.1007/978-3-642-34159-5_4. 2018). https://public.ccsds.org/Pubs/350x5g1.pdf.
[29] Colibri. 2014. PowerVu Management Keys Hacked. (De-
cember 5, 2014).
284