Large-Scale Measurement of NAT Traversal for the Decentralized Web: A Case Study of DCUtR in IPFS

Probably not a substantial claim: may not pass S3 (load-bearing).

Halkes et al. observed in 2011 that approximately 11% of their monitored peer cohort were behind EDM NAT devices.

“Halkes et al. [14] observed in 2011 that approxi- mately 11% of their monitored peer cohort were behind EDM NAT devices.”

• cites Halkes et al. [14] [14] Gertjan Halkes and Johan Pouwelse. 2011. UDP NAT and firewall puncturing in the wild. In Lecture Notes in Computer Science. Springer Berlin Heidelberg, Berlin, Heidelberg, 1–12. sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

NAT is a prevalent technique used in network gateways to modify IP address information, primarily to mitigate IPv4 address exhaustion.

“Network Address Translation (NAT) is a prevalent technique used in network gateways to modify IP address information; primarily, to mitigate IPv4 address exhaustion [9].”

• cites [9] [9] Kjeld Borch Egevang and Paul Francis. 1994. The IP Network Address Translator (NAT). RFC 1631. doi:10.17487/RFC1631 sourcePartly supported

  The Overview of NAT describes NAT as a router function installed at a stub domain’s exit point, and §3.3 says it modifies IP addresses. The Introduction identifies IP address depletion as a motivation. But the Conclusions (p. 8) call NAT a possible short-term solution whose suitability still required experimentation; the paper does not establish that NAT was prevalent. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim. May combine two claims.

Based on the EDM NAT prevalence reported by Halkes et al., we expect to encounter approximately 19.6% mixed EDM/EIM peer combinations and 1.2% combinations where both peers are behind EDM NATs.

“Based on the EDM NAT prevalence reported by Halkes et al. [14], we expect to encounter approximately 19.6% mixed EDM/EIM peer combinations and 1.2% combinations where both peers are behind EDM NATs.”

• cites Halkes et al. [14] [14] Gertjan Halkes and Johan Pouwelse. 2011. UDP NAT and firewall puncturing in the wild. In Lecture Notes in Computer Science. Springer Berlin Heidelberg, Berlin, Heidelberg, 1–12. sourceSupported

May combine two claims.

The ~80% QUIC share reflects only that QUIC's connection establishment latency is lower than TCP's, a finding consistent with Liang et al. [21], and does not directly speak to hole-punch success.

“This reflects only that QUIC's connection establishment latency is lower than TCP's, a finding consistent with Liang et al. [21], and does not directly speak to hole-punch success.”

• cites Liang et al. [21] [21] Jinyu Liang, Wei Xu, Taotao Wang, Qing Yang, and Shengli Zhang. 2024. Imple- menting NAT Hole Punching with QUIC. doi:10.48550/arXiv.2408.01791 sourceSupported

May combine two claims.

The data from Halkes et al. presents limitations due to its age (2011) and relatively small sample size (fewer than 2,000 peers).

“The data from Halkes et al. [14] presents limitations due to its age (2011) and relatively small sample size (fewer than 2,000 peers).”

• cites Halkes et al. [14] [14] Gertjan Halkes and Johan Pouwelse. 2011. UDP NAT and firewall puncturing in the wild. In Lecture Notes in Computer Science. Springer Berlin Heidelberg, Berlin, Heidelberg, 1–12. sourceSupported

May combine two claims.

This rate depends on successful relay reservation and public address discovery which themselves fail for approximately 29% of attempts.

“This rate depends on successful relay reservation and public address discovery which themselves fail for approximately 29% of attempts (see Section 6).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The number of contributed hole punches increased after December 1, declined over Christmas, and tapered off following the distribution of the campaign termination notice in early January.

“The graph shows that the number of contributed hole punches increased after December 1, declined over Christmas, and tapered off following the distribution of the campaign termination notice in early January.”

May combine two claims.

Applying the network identification method, we identify 859 distinct networks that the clients operate in (no two clients operated from the same network) and 86,769 networks for remote peers.

“Applying this method, we identify 859 distinct networks that the clients operate in (no two clients operated from the same network) and 86, 769 networks for remote peers.”

• AssumptionThe authors state this as an assumption: Data points belong to the same network if the client reports the same public IP address.
• AssumptionThe authors state this as an assumption: If a client's public IP changes but remains within the same Autonomous System (AS), the measurements are from the same network provided that the set of locally assigned private IP addresses remains unchanged.

May combine two claims.

In total, over 6.25 million hole punches were tracked across 212 API keys while 148 sign-ups were registered.

“In total, we tracked over 6.25 million hole punches across 212 API keys while registering 148 sign-ups.”

May combine two claims.

Approximately 10% of peers report a higher RTT after the hole punch, while 90% benefit from a reduced RTT.

“Approximately 10% of peers report a higher RTT after the hole punch (all points above 100), while 90% benefit from a reduced RTT.”

May combine two claims.

For critical relay-mediated RTTs (scenario 2), over 90% of measurements exhibit a standard deviation less than half of the average RTT, indicating a substantial synchronization margin.

“The results show that for the critical relay-mediated RTTs (scenario 2), over 90% of measurements exhibit a standard deviation less than half of the average RTT, indicating a substantial synchronization margin.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The sample distribution is skewed toward the U.S.

“While the sample distibution is skewed toward the U.S., our dataset captures globally distributed measurements.”

May combine two claims.

Clients were deployed in 39 countries and hole-punched remote peers in 167 countries.

“Figures 3a and 3b show that the clients were deployed in 39 countries and hole-punched remote peers in 167 countries.”

Probably not a substantial claim: may not pass S4 (based).

The central finding of transport-agnosticism of DCUtR is a result of the effective RTT-based synchronization.

“The viability of this roadmap is under- scored by the central finding of transport-agnosticism of DCUtR as a result of the effective RTT-based synchronization.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The remaining approximately 30% failure rate represents the next frontier for DCUtR improvement.

“The remaining ∼ 30% failure rate represents the next frontier for DCUtR improvement.”

May combine two claims.

Most relay nodes fall near the midpoint between us and the remote peer, with a slight skew toward proximity to the client.

“Most relay nodes fall near the midpoint between us and the remote peer, with a slight skew toward proximity to the client.”

May combine two claims.

Failed hole-punch attempts exhibit slightly higher RTTs through the relay than successful attempts, suggesting a weak negative correlation, but the overall difference is minor as the two CDFs remain closely aligned.

“While failed attempts exhibit slightly higher RTTs, suggesting a weak negative correlation, the overall difference is minor as the two CDFs remain closely aligned.”

Probably not a substantial claim: may not pass S3 (load-bearing).

One highly mobile client was observed operating from 28 distinct networks during the measurement period.

“At the other extreme, one highly mobile client was observed operating from 28 distinct networks during the measurement period.”

• AssumptionThe authors state this as an assumption: Data points belong to the same network if the client reports the same public IP address.
• AssumptionThe authors state this as an assumption: If a client's public IP changes but remains within the same Autonomous System (AS), the measurements are from the same network provided that the set of locally assigned private IP addresses remains unchanged.

May combine two claims.

Synchronization is highly effective, with most successes occurring on the first attempt (H3a).

“that synchronization is highly effective, with most successes occurring on the first attempt (H3a)”

Probably not a substantial claim: may not pass S3 (load-bearing).

Our CC BY-SA licensed dataset of over 4.4 million measurements across more than 85k networks globally provides a valuable resource for future research.

“Our CC BY-SA licensed dataset of over 4.4 million measurements across more than 85k networks globally provides a valuable resource for future research.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The provided dataset is a Postgres database dump available via a specific IPFS CID.

“The provided dataset is a Postgres database dump available via this IPFS CID”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Two conflicting trends have emerged since Halkes et al.'s study: (i) expected reduction in EDM NAT deployment as vendors comply with BEHAVE RFC, and (ii) proliferation of Carrier-Grade NAT (CGNAT) deployments, which employed EDM translation in 40% of cases as of 2016.

“Two conflicting trends have emerged since their study: (i) the ex- pected reduction in EDM NAT deployment as vendors increasingly comply with the BEHAVE RFC [17], and (ii) the proliferation of Carrier-Grade NAT (CGNAT) deployments, which employed EDM translation in 40% of cases as of 2016 [28].”

• cites [17] [17] Cullen Fluffy Jennings and Francois Audet. 2007. Network Address Translation (NAT) Behavioral Requirements for Unicast UDP. RFC 4787. doi:10.17487/ RFC4787 sourceNot checked — no full text of the source
• cites [28] [28] Philipp Richter, FlorianWohlfart, Narseo Vallina-Rodriguez, Mark Allman, Randy Bush, Anja Feldmann, Christian Kreibich, Nicholas Weaver, and Vern Paxson.Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

No abuse by non-registered clients was observed.

“Any detected abuse by non-registered clients could be excluded from the dataset; however, we did not observe any such behavior.”

Probably not a substantial claim: may not pass S3 (load-bearing).

NAPT allows multiple devices within a private network to share a single public IP address.

“The most common variant, Network Address Port Translation (NAPT), also known as Port Address Translation (PAT), allows multiple devices within a private network to share a single public IP address [15].”

• cites [15] [15] Matt Holdrege and Pyda Srisuresh. 1999. IP Network Address Translator (NAT) Terminology and Considerations. RFC 2663. doi:10.17487/RFC2663 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

DCUtR's permissionless design introduces several security trade-offs.

“DCUtR's permissionless design intro- duces several security trade-offs.”

Probably not a substantial claim: may not pass S2 (checkable).

If a direct connection can be established via a pre-existing port mapping, it is highly advantageous to prioritize this method over initiating a full hole punch.

“If a direct connection can be established via a pre-existing port mapping, it is highly advantageous to prioritize this method over initiating a full hole punch.”

Probably not a substantial claim: may not pass S3 (load-bearing).

This analysis is based on port-mapping reports from our 212 volunteer clients, whose network configurations may not be representative of the broader population.

“We note that this analysis is based on port-mapping reports from our 212 volunteer clients, whose network configurations may not be representative of the broader population.”

Classified as probably not a substantial claim.

Aggressive port allocation patterns may trigger network-level intrusion detection systems.

“Aggressive port allocation patterns may trigger network-level in- trusion detection systems.”

May combine two claims.

We cannot and do not attribute individual failures to specific NAT configurations.

“As a result, we cannot and do not attribute individual failures to specific NAT configurations.”

May combine two claims.

Our dataset does not support a definitive classification of remote NAT types, preventing attribution of individual failures to specific NAT mapping or filtering behaviors.

“our dataset does not support a definitive classification of remote NAT types, preventing us from attributing individual failures to specific NAT mapping or filtering behaviors”

May combine two claims.

Guha et al.'s measurements comprised just 93 home NATs, which offers limited statistical power and lower representativeness than our considerably larger dataset.

“However, their measurements comprised just 93 home NATs, which offers limited statistical power and lower representativeness than our considerably larger dataset.”

May combine two claims.

DCUtR's deterministic synchronization mechanism effectively mitigates the traditional challenges of TCP traversal, demonstrating that the protocol is transport-agnostic.

“DCUtR's deterministic synchronization mechanism effectively mitigates the traditional challenges of TCP traversal, demonstrating that the protocol is indeed transport-agnostic.”

Classified as probably not a substantial claim.

Opening a high volume of ports (e.g., 256-2,048) simultaneously increases a node's attack surface: each open port represents a potential entry point for fingerprinting or exploitation by adversaries.

“Beyond NAT resource limits, opening a high volume of ports (e.g., 256–2,048) simultaneously also increases a node's attack surface: each open port represents a potential entry point for fingerprinting or exploitation by adversaries who may be aware of the technique.”

Probably not a substantial claim: may not pass S3 (load-bearing).

This pattern aligns with the timing of our study and expected client availability during the holiday period.

“This pattern aligns with the timing of our study and expected client availability during the holiday period (see Figures 3a and 3b).”

Probably not a substantial claim: may not pass S3 (load-bearing).

TCP hole punching is widely considered difficult due to constraints of the three-way handshake and susceptibility to RST packets from stateful firewalls.

“TCP hole punching is widely considered difficult due to constraints of the three-way handshake and susceptibility to RST packets from stateful firewalls, which are well-documented [11, 13].”

• cites [11] [11] Bryan Ford, Dan Kegel, and Pyda Srisuresh. 2008. State of Peer-to-Peer (P2P) Communication across Network Address Translators (NATs). RFC 5128. doi:10. 17487/RFC5128Not checked yet

  Not checked in the free run.

• cites [13] [13] Saikat Guha and Paul Francis. 2005. Characterization and measurement of TCP traversal through NATs and firewalls. In Proceedings of the 5th ACM SIGCOMM Conference on Internet Measurement (Berkeley, CA) (IMC '05). USENIX Associa- tion, USA, 18. sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Failures can arise from transient network conditions, misconfigurations, overly restrictive firewalls, or the prevalence of Symmetric NATs.

“Failures can arise from transient network conditions, misconfigurations, overly restrictive firewalls, or the prevalence of Symmetric NATs.”

Probably not a substantial claim: may not pass S3 (load-bearing).

When the search space coverage increases to 3.1% (2,048 ports), the success probability rises to 99.9%.

“When the search space coverage increases to 3.1% (2,048 ports), the success probabil- ity rises to 99.9%.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The relay represents a trust boundary and can observe the addresses and connection patterns of both parties, enabling behavioral profiling over time.

“The relay itself represents a trust boundary. It can observe the addresses and connection patterns of both parties, enabling behavioral profiling over time.”

Classified as probably not a substantial claim.

The 1,000 data point threshold ensures statistical stability of per-network success rates by excluding transient or minimally contributing clients whose small sample sizes would introduce high variance into the aggregate metric.

“The 1,000 data point threshold ensures statistical stability of per-network success rates by excluding transient or minimally contributing clients whose small sample sizes would introduce high variance into the aggregate metric.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Symmetric NATs pose the greatest challenge for P2P connectivity.

“At the opposite end of the spectrum are Symmetric NATs, which pose the greatest challenge for P2P connectivity.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The excluded data points represent fundamental failures in the broader connectivity stack upon which DCUtR depends.

“The excluded data points represent fundamental failures in the broader connectivity stack upon which DCUtR depends.”

May combine two claims.

TURN guarantees connectivity but increases latency and requires the TURN server to handle the full bandwidth of the session.

“Although TURN guarantees connectivity, it increases latency and requires the TURN server to handle the full bandwidth of the session.”

Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

If birthday paradox techniques achieve a 64% success rate for the 19.6% of mixed connection attempts, the overall DCUtR success rate could improve 12.5%.

“If birthday paradox techniques achieve a 64% success rate for the 19.6% of mixed connection attempts, the overall DCUtR success rate could improve 12.5%.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Repeated failed TCP simultaneous-open attempts can be misidentified as SYN floods by stateful firewalls, risking IP denylisting, while resource-constrained NATs may exhaust their session tables under sustained probing.

“On the denial-of-service front, re- peated failed TCP simultaneous-open attempts can be misidentified as SYN floods by stateful firewalls, risking IP denylisting, while resource-constrained NATs may exhaust their session tables under sustained probing.”

Classified as probably not a substantial claim.

The approximately 30% failure rate, despite the protocol's general efficacy (H1), is likely due, in part, to challenging NAT configurations, such as Endpoint-Dependent Mappings (EDM).

“The ∼ 30% failure rate, despite the protocol's general efficacy (H1), is likely due, in part, to challenging NAT configurations, such as Endpoint-Dependent Mappings (EDM).”

May combine two claims.

When path asymmetry is significant, the synchronization margin shrinks, which may partly explain the remaining ~30% failure rate.

“When path asymmetry is significant, however, the synchronization margin shrinks, which may partly explain the remaining ∼ 30% failure rate.”

May combine two claims.

Taken together, these findings demonstrate only a minor effect from elevated RTTs and no measurable dependence on relay path location, providing compelling evidence for H2.

“Taken together, these findings demonstrate only a minor effect from elevated RTTs and no measurable dependence on relay path location, providing compelling evidence for H2.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Circuit v2 is designed to minimize resource usage, imposing negligible processing and bandwidth costs on relays, unlike traditional TURN relays.

“Crucially, Circuit v2 is designed to minimize resource usage, imposing negligible processing and bandwidth costs on relays; unlike traditional TURN relays, which forwards all application traffic.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Without the conservative constraints on port probing and opening operations, success rates could be substantially higher.

“Without these constraints, success rates could be substantially higher.”

• AssumptionThe authors state this as an assumption: Conservative values for port probing and opening operations are assumed because NAT devices maintain limited active session capacities and may interpret intensive port scanning as malicious behavior.

May combine two claims.

The skew toward proximity to the client is expected because peers by default request reservations with two relays, clients dial remote peers via all available relay addresses, and the connection to the closer relay succeeds first.

“This skew is expected, since peers by default request reservations with two relays. Clients then dial remote peers via all available relay addresses. It is natural that the connection to the closer relay, and thus to the remote peer through it, succeeds first.”

• AssumptionRests on an assumption: Peers by default request reservations with two relays.

Classified as probably not a substantial claim.

This insight motivates the optimizations described in Section 7.

“This insight motivates the optimizations described in Section 7.”

Probably not a substantial claim: may not pass S3 (load-bearing).

ICE is the de facto standard for handling the complexities of NAT.

“The de facto standard for handling the complexities of NAT is the Interactive Connectivity Establishment (ICE) framework [1, 29]”

• cites [1] [1] Harald T. Alvestrand. 2021. Overview: Real-Time Protocols for Browser-Based Applications. RFC 8825. doi:10.17487/RFC8825 sourceNot checked — no full text of the source
• cites [29]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Personally identifiable information has been anonymized.

“Personally identifiable information has been anonymized.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Cases where the direct path is slower than the relayed path can occur if the relay is located on a high-speed internet backbone, while the direct peer-to-peer path traverses slower consumer-grade ISP networks.

“Cases where the direct path is slower than the relayed path can occur if the relay is located on a high-speed internet backbone, while the direct peer-to-peer path traverses slower consumer-grade ISP networks.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The Identify protocol provides functionality similar to STUN, is lightweight, and all libp2p peers support it by default.

“The Identify protocol provides functionality similar to STUN, is lightweight, and all libp2p peers support it by default.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

All studies we identified rely on data that is considerably outdated, which may not accurately reflect the current internet landscape and underscores the need for renewed, large-scale measurement of NAT traversal in the current Internet.

“Moreover, all studies we identified rely on data that is considerably outdated. This may not accurately reflect the current internet landscape and underscores the need for renewed, large-scale measurement of NAT traversal in the current Internet.”

May combine two claims.

This selection bias may lead to an overestimation of the protocol's success rate (H1) and may limit the generalizability of our findings.

“This selection bias may lead to an overestimation of the protocol's success rate (H1) and may limit the generalizability of our findings.”

May combine two claims.

Hole punching can significantly improve performance and enable delay-sensitive applications.

“These results confirm that hole punching can significantly improve performance and enable delay-sensitive applications.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The research followed strict ethical guidelines.

“The research followed strict ethical guidelines.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The Go-Client generated ten peer identities upon startup, each listening on a different port, and for each hole punch we round-robin through the ten peers, mitigating persistent port mappings to yield a high number of CONNECTION_REVERSED results.

“Importantly, the Go-Client generated ten peer identities upon startup. Each peer ID would listen on a different port. For each hole punch we 'round robin' through the ten peers. This way we mitigated persistent port mappings to yield a high number of CONNECTION_REVERSED results.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

The proposed optimizations of role alternation on retries, refined RTT calculation, proactive NAT priming, and the birthday paradox exploitation offer concrete paths toward higher direct-connectivity rates and strengthen the protocol as a building block for the decentralized Internet.

“The proposed optimizations of role alter- nation on retries, refined RTT calculation, proactive NAT priming, and the birthday paradox exploitation offer concrete paths toward higher direct-connectivity rates and strengthen the protocol as a building block for the decentralized Internet.”

The quote only partly states this claim.

Data from remote, non-consenting peers was collected based on their operation of public, permissionless P2P nodes (IPFS, libp2p), which inherently expose network metadata (IP, PeerID, protocols) through standard interactions.

“Data from remote, non-consenting peers was collected based on their operation of public, permissionless P2P nodes (IPFS, libp2p), which inherently expose network metadata (IP, PeerID, protocols) through standard interactions.”

May combine two claims.

Participants either signed up, receiving detailed information and providing explicit consent regarding data collection (including IP addresses and libp2p PeerIDs) and its academic use, or voluntarily ran a tool whose documentation clearly stated data collection for research.

“Participants either signed up, receiving detailed information and providing explicit consent regarding data collection (including IP addresses and libp2p PeerIDs) and its academic use, or voluntarily ran a tool whose documentation (README) clearly stated data collection (including IP addresses and libp2p PeerIDs) for research.”

May combine two claims.

Collected data (e.g., hole punch success) was a byproduct of standard interactions, with no non-standard or privacy-invasive data gathered beyond what a public P2P node inherently exposes.

“Collected data (e.g., hole punch success) was a byproduct of these standard interactions, with no non-standard or privacy-invasive data gathered beyond what a public P2P node inherently exposes.”

May combine two claims.

While not a full replacement for STUN/TURN today, DCUtR can already offload significant traffic and, when combined with suitable relays (e.g. VPN or SSH gateways), has the potential to eliminate dependence on these services entirely.

“While not a full replacement for STUN/TURN today, it can already offload significant traffic and, when combined with suitable relays (e.g. VPN or SSH gateways), has the potential to eliminate dependence on these services entirely.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Alternating roles in subsequent attempts would succeed on the second try, given that the initiator is not behind a symmetric NAT.

“The above scenario would succeed on the second try, given that the initiator is not behind a symmetric NAT.”

• AssumptionThe authors state this as an assumption: The initiator is behind an EDM NAT but the listener is not.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Having both parties sending client hellos might result in two distinct QUIC connections, which can also be acceptable in some scenarios.

“Contrary to a TCP simultaneous open, having both parties sending client hellos might result in two distinct QUIC connections, which can also be acceptable in some scenarios.”

May combine two claims.

DCUtR meaningfully advances decentralization by reducing reliance on centralized intermediaries, thus mitigating single points of failure and enhancing self-sovereign control over data.

“DCUtR meaningfully advances decentralization by reducing re- liance on centralized intermediaries, thus mitigating single points of failure and enhancing self-sovereign control over data.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Long-term success also depends on improving the network environment itself through a combination of standards advocacy and better use of existing mechanisms.

“Long-term success also depends on improving the network environ- ment itself through a combination of standards advocacy and better use of existingmechanisms.”

Classified as probably not a substantial claim. May combine two claims.

Peer-to-peer RTT may be insufficient in more complex network topologies, where the resulting discrepancy can cause the SYNC packet to arrive outside the optimal time window, increasing the likelihood of connection failures.

“this approach may be insufficient in more complex network topologies. In such cases, the resulting discrepancy can cause the SYNC packet to arrive outside the optimal time window, thereby increasing the likeli- hood of connection failures.”

May combine two claims.

This paper presents a large-scale empirical evaluation of DCUtR, validating its design principles and establishing a contemporary benchmark for decentralized NAT traversal.

“This paper presents a large-scale empirical evaluation of DCUtR, validating its design principles and establishing a contemporary benchmark for decentralized NAT traversal.”