Associate prof and CS honors chair @umdcs, Security researcher. Founder of breakerspace.io
College Park, MD
- Tweets 2,436
- Following 297
- Followers 2,407
- Likes 3,286
My Dad unexpectedly passed away last Thursday at the age of 84, peacefully, calmly, in the house I grew up in. He was an inventor, a radio operator, a pilot, an entrepreneur, an amazing father and pop-pop. He never let a loving thought go unsaid. Love you, pop.
ALT Dad and Dave at Christmas brunch
ALT Dad with a beaming smile at Dave's wedding
3
13
624
On @Cloudflare Radar, you can now track various forms of traffic tampering around the world! This is the result of our collaboration between @UofMaryland and Cloudflare to develop passive ways of detecting when third parties tamper with user traffic. Stay tuned for a blog post!
14
2,477
Oh that was my broader call for more data (relevant picture attached 😊)-- I'm not trying to dump this on you!
1
69
I bought this shirt ~15 years ago and waited for the right opportunity to wear it. Today was that day, as I performed a science experiment with my kid's class. Thank you @PHDcomics for helping introduce a love for data to a new generation!
18
969
There are many other instances of crime and re-victimization in the paper. Here's a table summarizing 15 of the phones.
1
375
We found many instances of identity theft, including victims' social social security numbers, credit card numbers, bank account information, passports and other forms of identification—the literal definitions of "personally identifiable information".
1
338
First, there were all the things you might expect: text messages, photos (including nudity, drugs, weapons), browsing histories, cookies, usernames, passwords,...
1
344
That was not the only time a phone's credentials were included in the auction. Here is another PropertyRoom auction from Jan 2020 that has what appears to be a swipe pattern on it (we did not purchase this phone or confirm).
1
354
The police appear to have cracked one phone's PIN using @GrayshiftLLC's GrayKey. Here's the picture from PropertyRoom. This phone arrived with a sticky-note on the back; we confirmed the number was the PIN. This was not one of the most common PINs; we would not have guessed it.
1
1
3
1,552
Of the 228 we purchased, we got full access to 49 of them JUST BY TURNING THEM ON. They had no passcode; they arrived completely unlocked. We tried the 100 most common PINs/patterns ("1234", etc.) and that got us access to another 11 phones.
1
601
Whatever the group needs to collectively achieve their goals, in terms of both productivity and happiness. Sometimes that'll mean a non-coding manager, sometimes a coding manager. Blanket rules never work.
477
Reviewers at the birth of their child:
“Ten fingers and ten toes. Methodology lacks novelty. Reject.”
5
14
1,753
To whatever reviewer wrote this, it feels really good to be seen:
nice to see a paper all done with gnuplot! 👏
1
9
1,215
At the time, ~15% of South America reassigned within 12 hours. We found in a paper a few years ago that Brazil had more Hajime bots than any other country, by far. That said, CARDCount will still be useful for you when you want to count bots over longer periods of time!
1
1
65
Paper, code, and data to come. Stay tuned! But if you have a dataset of IP addresses and want to count them, or if you just want to understand a botnet better, reach out!
Congrats again to @_LeonBock and fellow authors @rameses888 @chr_doerr and Max Mühlhäuser!
3
1
4
872
We call it CARDCount: Considering Address Reassignment Durations when counting. where is other techniques only let you count, bots in small, sub our snap shots, CARDCount gives accurate estimates over weeks.
1
3
348
The high-level idea is to learn how long ISPs assign IP addresses then use that to determine if two IP addresses are likely from the same host. If an ISP reassigns every 24 hours and you see 4 IPs over 4 days, it’s probably one bot. Sounds simple, but the devil’s in the details!
1
1
350
We all know that IP addresses are poor long-term identifiers, because they’re not permanently assigned. But what if you want to accurately count the bots in a botnet and all you have is IP addresses? Thanks to @_LeonBock’s first @NDSSSymposium paper we have a solution! 🧵
3
10
26
5,076
Police: “get off the ice or we will arrest you!”
Police later that day: “get onto the ice and get the furniture or we will arrest you!”
1
282
Senator Booker thinks that because this is a “nonviolent” crime so her punishment should be “fair and proportionate” (i.e., shortened). I would much rather be the victim of a violent crime than terminate a pregnancy that Theranos knowingly incorrectly said was not viable.
1
1
Recorded a podcast as a guest on the @packetpushers! We covered it all: research, open access publications, AI, and the competitive world of tuba auditions. The episode should drop late this month — stay tuned!
1
1
21
Come find me over on Mastodon: @DistributedDave@infosec.exchange — if anyone else has any recommendations for servers, lmk! Where are you?
2
Made it to school in time to introduce @LeoLambro to the CS Honors students, so he can tell them all about his research finding bugs through fuzzing!
1
4
Taking advantage of the nice weather to run to @UofMaryland from my home in the heart of @washingtondc. Apply to UMD, and enjoy being so close to the nation’s capital.
2
1
30
Thrilled to have @jbhuang0604 presenting his work—and the history of humanity leading up to this selfie—to @umdcs undergraduate honors students!
1
3
The papers I linked above describe half a dozen server-side strategies that evade censorship in Iran. Some might work for you; if they do, let us know! If you want to try, let us know! And if you want to learn more, check out censorship.ai
2
1
3
Two years ago, Iran turned on their protocol filter: an extra layer of censorship that permits only a limited set of protocols. Within about an hour, Geneva found server-side strategies to evade the filter and its traditional censorship infrastructure.
cs.umd.edu/~dml/papers/iran_…
1
3
Geneva discovered the first server-side censorship evasion strategies that work with totally unmodified clients. The server does all the work for the client; this helps people who didn’t even know they were being censored in the first place.
cs.umd.edu/~dml/papers/genev…
1
4
This tool, Geneva, trains directly against live censors to discover new sequences of packets that evade censorship without adversely affecting the service the client is trying to reach. Training against real censors, it takes advantage both of design flaws and bugs.
1
2
Is your website being censored in Iran? We might be able to help. We have developed an AI-based tool that automatically learned how to evade censorship without requiring clients inside Iran to download any extra software whatsoever 🧵
1
3
13
It appears to be a bit more nuanced with SC retailers. But what’s especially weird to me is that this is an example of the exception that proves the rule: by saying they can’t sell to under-21s 7pm-9am they are implying that they CAN sell to them at the other times of day 🤔
1
