Tor Stinks: Stealthy Traffic Analysis
Stealthy Traffic Analysis of Low-Latency Anonymous Communication Using Throughput Fingerprinting Prateek Mittal et al. ACM CCS 2011 https://dl.acm.org/doi/pdf/10.1145/2046707.2046732 https://www.cryptolux.org/images/b/bc/Tor_Issues_Thesis_Thill_Fabrice.pdf https://ohmygodel.com/publications/peerflow-popets2017.pdf
On Wed, Sep 16, 2020 at 05:54:17PM -0400, grarpamp wrote:
Thanks for posting this link.
https://www.hackerfactor.com/blog/index.php?/archives/897-Tor-0day-Crashing-...
circuit padding client-edge padding capability already built
Neither may have much impact on a global TA adversary. Draw it on paper and start thinking like a TA, if not obvious, then drop in a cut. As before on this list... a full time base layer of fill among connected p2p nodes seems more potential for impact against TA.
someone said to tack up another N x parallel connections between the same two nodes when more bandwith desired
Such channel bonding, socket / cpu / management resources... likely N x inefficient. Renegotiate the existing connection, design it as a cell carrier from start, etc.
Hey, has anyone else ever implemented this before, and maybe had a patch rejected?
Tor eschews entertaining such models and options on their lists as Tor dislikes traffic and will not accept such a patch that alters its marketed mode of operation. Thus tor will remain 25+ year old TA vulnerable system, leaving new projects to explore various approaches to TA in a low latency net (and even make cost raising impact on Sybil). Some coders who know Tor Project is a censor rejecting topical posts and critiques, lies to its users, is infected with virtue signalling SJW, etc... might not feel comfy working them with patches either, as opposed to with a fork or some new. Edge padding, and recent not-enough-against-anti-sybil, moved forward, in some part because small number noisy people on TA/Sybil, who were more banned than noise/ideas thanked.
more than
After people made suggestions years ago, Tor is only now finally getting around to finding 300 bad relays... https://lists.torproject.org/pipermail/tor-relays/2020-October/019045.html protip: There are hundreds more than that...
Another 600++ were removed in September. Estimate 25-50% of all nodes are adversarial. And the network is not very TA resistant. With darknet markets constantly getting shut. Tor Project still advertising false protection and not putting up giant warning banners.
Four Seasons Total Interior Design https://twitter.com/mattblaze/status/1332442077309833217
Matt Blaze, who is Chair of board of directors of Tor Project Inc, and supposedly a respected security analyst, appears to spend more time twatting about stupid microphones, than removing the needless rampant censorship within Tor Project, or correcting the security distortions shilled to users on Tor's front page. Perhaps that's two other embarassing facts that Tor Project doesn't want to be discussed on it's comms channels. But hey, at least Tor is trendy hipster partners with the Nanny State, it now officially tells you to "Wear a Mask" right at the top of its front page, along with "Resist the Pandemic". A new State design for all seasons indeed.
HF posted:
https://www.hackerfactor.com/blog/index.php?/categories/19-Tor https://twitter.com/hackerfactor/status/1341164309095694336 "my years of previous interactions with the Tor Project had been met with everything from silence to outright hostility. ... With no vendor response, I escalated to a blog series about Tor's vulnerabilities. ... The Tor Project went on the defensive, trying to mitigate the risks called out in my previous two blog entries. ... For the goal of getting them to fix something, anything, this project failed. However, I ended up learning much more about the Tor Project... -- HF"
grarpamp posted:
Another 600++ were removed in September. Estimate 25-50% of all nodes are adversarial.
JDB posted:
Over 25% Of Tor Exit Relays Are Spying On Users' Dark Web Activities https://thehackernews.com/2021/05/over-25-of-tor-exit-relays-are-spying.html
https://nusenu.medium.com/tracking-one-year-of-malicious-tor-exit-relay-acti... Tor network has been infested with bad nodes for ~20 years ever since inception [1]. Tor Project and "community" were suggested many times to create some new userland mindsets on risk, independent outside node analysis groups and node cooperations, new available options for users and usage... allowing users to plugin various externally maintained node selection metrics and subscriptions including based on creating various WoT's of nodes. Tor Project quietly never bothered to make discuss or help promote those distributed assertion/usage models further. Instead they still limit scope to central blocking now 1000 of obviously malicious bad nodes a month, while leaving users hanging subjected to a large percent of undetect[ed/able] bad nodes (including TA middles/guards modulators) many which been around for nearly as many years, and making press releases about how their blocking is keeping users safe. Are now pushing contact field labeling as non-solution that do nothing (because the malactors do the labeling, sheep meet wolf, lol). Tor Project still takes quiet inflection that it protects users, advertises tor is stronger than it is, and users should accept and use their pontifically safe design and relays on faith, and not raise counterpoint question critique expo[s/$]e or options for user benefit, else posters be censored from Tor comms channels and people kicked from project and areas. Plus extra boot for those maintaining independence from wokeism. "Tor Stinks -- NSA" [1] Just a fraction of all the nodes blocked this month alone... https://lists.torproject.org/pipermail/tor-relays/2021-May/019644.html Another partial list... https://gitlab.torproject.org/tpo/network-health/team/-/wikis/Rejected-finge... New overlay network designs and operations need to arise, and from places far from involved in Tor Project.
Speaking of all that... The fact that "Tor Stinks -- NSA"... the need to start fresh investigation, code, develop, and deploy additional new alternative proof-of-concept networks, besides the old 25 year legacy and vacuum that is Tor Project, utilizing full-time base of chaff fill, and other old and new methods that have not hardly well considered and operationally proof tested since yet... including a fresh investigation of mix based networks, is obvious, and is slowly being corroborated, and has been attested by both info-theory and classified operations of the past 70 years... Emphasis * quotes added... https://en.wikipedia.org/wiki/One-time_pad During World War II and into the 1950s, the U.S. made extensive use of one-time tape systems. In addition to providing confidentiality, *****circuits secured by one-time tape ran continually, *even when there was no traffic*, thus protecting against traffic analysis.***** In 1955, NSA produced some 1,660,000 rolls of one time tape. Each roll was 8 inches in diameter, contained 100,000 characters, lasted 166 minutes and cost $4.55 to produce. https://en.wikipedia.org/wiki/KW-26 It was developed in the 1950s by the National Security Agency (NSA) to secure fixed teleprinter circuits that **operated 24 hours a day.** Because the KW-26 sent **a continuous stream of bits, it offered traffic-flow security**. Someone intercepting the ciphertext stream **had no way to judge how many real messages were being sent, making traffic analysis impossible.** https://en.wikipedia.org/wiki/Traffic_analysis " In computer security Traffic analysis is also a concern in computer security. An attacker can gain important information by monitoring the frequency and timing of network packets. **A timing attack on the SSH protocol can use timing information to deduce information** about passwords since, during interactive session, SSH transmits each keystroke as a message.[8] **The time between keystroke messages can be studied** using hidden Markov models. Song, et al. claim that it can recover the password fifty times faster than a brute force attack. Onion routing systems are used to gain anonymity. *****Traffic analysis can be used to attack anonymous communication systems like the Tor anonymity network.***** Adam Back, Ulf Möeller and Anton Stiglic present traffic analysis attacks against anonymity providing systems .[9] Steven J. Murdoch and George Danezis from University of Cambridge presented [10] research showing that ***traffic-analysis allows adversaries to infer which nodes relay the anonymous streams.*** This reduces the anonymity provided by Tor. They have shown that otherwise unrelated streams can be linked back to the same initiator. Remailer systems can also be attacked via traffic analysis. **If a message is observed going to a remailing server, and an identical-length (if now anonymized) message is seen exiting the server soon after, a traffic analyst may be able to (automatically) connect the sender with the ultimate receiver.** Variations of remailer operations exist that can make traffic analysis less effective. Countermeasures It is difficult to defeat traffic analysis without both encrypting messages and masking the channel. ***When no actual messages are being sent, the channel can be masked [11] *by sending dummy traffic*, similar to the encrypted traffic, thereby keeping bandwidth usage constant*** .[12] "It is very hard to hide information about the size or timing of messages. ***The known solutions require Alice to send a *continuous stream* of messages at the maximum bandwidth she will ever use.*** Song, Dawn Xiaodong; Wagner, David; Tian, Xuqing (2001). "Timing Analysis of Keystrokes and Timing Attacks on SSH". 10th USENIX Security Symposium. http://www.cypherspace.org/adam/pubs/traffic.pdf http://www.cl.cam.ac.uk/users/sjm217/papers/oakland05torta.pdf https://web.archive.org/web/20060913152709/http://students.cs.tamu.edu/xinwe... " Tor's new "netflow analysis resistance" is limited applicability, not really a general case attempt at diminishing the global/p2p/e2e/n2n Traffic Analysis problem by at least say a worthwhile factor of 10. At least one outline of how to consider doing a base of full time dynamically yielding chaff fill has already been posted at times to these lists. And there are plenty more academic papers on the subject proffering other methods to consider as well. Plus new ones that will come from that process. The user experience under a dynamic chaff and allocated bandwidth system is not likely to be as scarily unusable as some legacy entrenched project$ often like to claim. Timed buckets carrying stuff between nodes worked for ATM networks in the telcos, so for one investigation, see what kind of performance you can get emulating with today's sw, cpu's, and nic's over the net. https://www.hackerfactor.com/blog/index.php?/categories/19-Tor "Tor Stinks -- NSA"
https://www.vice.com/en/article/jg84yy/data-brokers-netflow-data-team-cymru https://docs.paloaltonetworks.com/cortex/cortex-xpanse/cortex-xpanse-user-gu... https://citizenlab.ca/2021/07/hooking-candiru-another-mercenary-spyware-vend... https://www.vice.com/en/article/88ng8x/pentagon-americans-surveillance-witho... https://www.vice.com/en/article/qj454d/private-intelligence-location-data-xm... https://www.vice.com/en/article/jgqm5x/us-military-location-data-xmode-locat... ISPs are quietly distributing "netflow" data that can, among other things, trace traffic through VPNs. There's something of an open secret in the cybersecurity world: internet service providers quietly give away detailed information about which computer is communicating with another to private businesses, which then sells access to that data to a range of third parties, according to multiple sources in the threat intelligence industry. The information, known as netflow data, is a useful tool for digital investigators. They can use it to identify servers being used by hackers, or to follow data as it is stolen. But the sale of this information still makes some people nervous because they are concerned about whose hands it may fall into. "I'm concerned that netflow data being offered for commercial purposes is a path to a dark fucking place," one source familiar with the data told Motherboard. Motherboard granted multiple sources anonymity to speak more candidly about industry issues. How Data Brokers Sell Access to the Backbone of the Internet ISPs are quietly distributing "netflow" data that can, among other things, trace traffic through VPNs. Joseph Cox by Joseph Cox August 24, 2021, 1:31pm Server Image: Cathryn Virginia/Motherboard Screen Shot 2021-02-24 at 3 Hacking. Disinformation. Surveillance. CYBER is Motherboard's podcast and reporting on the dark underbelly of the internet. See More → There's something of an open secret in the cybersecurity world: internet service providers quietly give away detailed information about which computer is communicating with another to private businesses, which then sells access to that data to a range of third parties, according to multiple sources in the threat intelligence industry. The information, known as netflow data, is a useful tool for digital investigators. They can use it to identify servers being used by hackers, or to follow data as it is stolen. But the sale of this information still makes some people nervous because they are concerned about whose hands it may fall into. "I'm concerned that netflow data being offered for commercial purposes is a path to a dark fucking place," one source familiar with the data told Motherboard. Motherboard granted multiple sources anonymity to speak more candidly about industry issues. At a high level, netflow data creates a picture of traffic flow and volume across a network. It can show which server communicated with another, information that may ordinarily only be available to the server owner or the ISP carrying the traffic. Crucially, this data can be used for, among other things, tracking traffic through virtual private networks, which are used to mask where someone is connecting to a server from, and by extension, their approximate physical location. Team Cymru, one threat intelligence firm, works with ISPs to access that netflow data, three sources said. Keith Chu, communications director for the office of Senator Ron Wyden which has been conducting its own investigations into the sale of sensitive data, added that Team Cymru told the office "it obtains netflow data from third parties in exchange for threat intelligence." Do you work at a company that handles netflow data? Do you work at an ISP distributing that data? Or do you know anything else about the trade of netflow data? We'd love to hear from you. Using a non-work phone or computer, you can contact Joseph Cox securely on Signal on +44 20 8133 5190, Wickr on josephcox, or email joseph.cox@vice.com. Companies that may source Team Cymru's data include cybersecurity firms hired to respond to data breaches or proactively hunt out hackers. On its website, Team Cymru says it works with both public and private sector teams to "to help identify, track and stop bad actors both in cyber space and on the ground." "I'm less worried about a bad guy hacker and more worried about a bad guy government or company or politician," one source familiar with the data said. A source in the threat intelligence industry added that they "always thought it was kinda bonkers," referring to Team Cymru's sale of netflow data. The continued sale of sensitive data could present its own privacy and security concerns, and the news highlights that ISPs are providing this data at scale to third parties likely without the informed consent of their own users. Other companies, such as cybersecurity firm Palo Alto Networks, also have access to netflow data. "The users almost certainly don't [know]" their data is being provided to Team Cymru, who then sells access to it, the source familiar with the data said. Tech Private Intel Firm Buys Location Data to Track People to their 'Doorstep' Joseph Cox 09.02.20 Team Cymru's customers can probe a dataset, and "effectively run queries against virtually any IP to pull the netflows to and from that IP over a given point in time," one of the sources said. Chu added Team Cymru said it "restricts the amount of data that is returned, so that only a small portion of the netflow data in its database can be accessed by any one client." In product descriptions, Team Cymru offers users the ability to follow traffic through VPNs, which attackers may use to cover their tracks or ordinary people to browse the internet more privately. "Trace malicious activity through a dozen or more proxies and VPNs to identify the origin of a cyber threat," one brochure for a Team Cymru product called Pure Signal Recon reads. In essence, access to netflow data lets a security team observe what is happening on the wider internet, and may indicate what is happening to other organizations, beyond the borders of their own network or company. One of the sources said they previously saw traffic from an organization they knew inside Team Cymru's dataset and was spooked by it at the time. "Visibility and insight are global," the description adds. An image included in the brochure shows Team Cymru's product letting users trace the activity of servers linked to an Iranian hacking group further than other datasets, such as DNS lookups. team-cymru-marketing.png A section of Team Cymru's marketing material for its Pure Signal Recon product. Image: Team Cymru. In a recent research report on an Israeli spyware vendor called Candiru, Citizen Lab thanked Team Cymru. "Thanks to Team Cymru for providing access to their Pure Signal Recon product. Their tool’s ability to show Internet traffic telemetry from the past three months provided the breakthrough we needed to identify the initial victim from Candiru’s infrastructure," the report reads. Citizen Lab did not respond to multiple requests for comment. Team Cymru did not respond to multiple requests for comment on which ISPs provide it with the data, what privacy protections are in place around the collection and distribution of such data, and whether the individual ISP users have provided consent for their data to be shared. "Fundamentally, people have a right to some degree of anonymity, and as a carrier it's not our job to eavesdrop in any form." For its Cortex Xpanse product, Palo Alto Networks also gains access to netflow data, according to product documentation available online. "Cortex® Xpanse™ obtains flow data via multiple relationships with Tier 1 ISPs. Through these relationships, Cortex Xpanse has access to a sample of approximately 80% of global flows," one page reads. Jim Finkle, director of threat communications at Palo Alto Networks, said in an emailed statement that "Palo Alto Networks provides enterprise customers with netflow data to and from their own networks to identify violations of security policies, gaps in security monitoring and other high-risk activity on the customer’s network." Palo Alto Networks declined to name which ISPs it sources data from, or whether it purchases the data outright from the ISPs. Dave Schaeffer, CEO of ISP Cogent Communications, which he said handles around 22 percent of the world's internet traffic, told Motherboard that as an ISP his company doesn't provide their netflow data to anybody. "Fundamentally, people have a right to some degree of anonymity, and as a carrier it's not our job to eavesdrop in any form," he said in a phone call. Schaeffer says Cogent generates 96 percent of its traffic from selling to large wholesale customers, such as Vodafone, Cox, Spectrum, and BT. Schaeffer says Cogent provides services to Team Cymru but does not share netflow data with the company. "I don't know if there's a lot of really useful things people could do with [netflow] data," he added. "There's probably some bad things I could think of if that data was available." Although multiple sources were concerned about the sale of netflow data, several of them stressed that Team Cymru is a responsible organization. "It's pretty shadowy but honestly they're a 'good actor,'" one in the threat intelligence industry said. "Very strict protections on who can see it, but still, yeah, it's shady." The source familiar with the data said they were concerned about the sale of netflow data, but that Team Cymru "also enable security organizations to do some really awesome work. So I'm conflicted about it." "I'm concerned that netflow data being offered for commercial purposes is a path to a dark fucking place." In May, Motherboard reported that Senator Wyden's office asked the Department of Defense (DoD), which includes various military and intelligence agencies such as the National Security Agency (NSA) and the Defense Intelligence Agency (DIA), for detailed information on its data purchasing practices. The response showed that the Pentagon is carrying out warrantless surveillance of Americans, according to a subsequent letter written by Wyden and obtained by Motherboard. Some of the answers the DoD provided were provided in a form meaning that Wyden's office could not legally publish specifics on the surveillance. Wyden's office then asked the DoD to release the information to the public. At the time, Wyden's office declined to provide Motherboard with specifics on one of the answers which was classified, but a Wyden aide said that the question related to the DoD buying internet metadata. "Are any DoD components buying and using without a court order internet metadata, including 'netflow' and Domain Name System (DNS) records," the question read. Other cybersecurity firms sell access to controversial datasets. In September, Motherboard reported how one firm called HYAS was sourcing smartphone location data to trace people to their "doorstep." As Motherboard has repeatedly shown, the ordinary apps installed on peoples' phones that gather this information often don't have informed consent to then sell or otherwise provide it to third parties.
participants (2)
-
grarpamp
-
Zenaan Harkness