I agree that a single word, either untouched or mangled, is not secure. This is true regardless of what language is used. Back in the 90's I used to use Hungarian words thinking that language was fairly obscure, only 11 million speak it. Wrong, it is now one of the standard languages used for dictionary attacks.
The current state of password cracking allows secure use of passphrases though. A coarse attack against a passphrase using a dictionary of 20,000 words requires an effort of 20,000 ^ N, where N is the number of words. If N is 2 that effort is 400,000,000 (actually 200 million on average) This is still not secure. 3 words requires an average effort of 4 trillion guesses. This is still not secure particularly if the words are not random but a sentence fragment. The security can be increased with mangling but it is better to choose a basic length without considering mangling. 5 random words requires and average effort of 1.6 x 10 ^ 21. This is very roughly equivalent to a binary key of length 70 (70 bits of entropy) and with mangling can approach a password that, depending on the hashing algorithm employed, even the NSA will, currently, have a hard time cracking in a reasonable time. If one uses a larger dictionary, say 1 million words including all sorts of technical words, a random 3 word passphrase requires an effort of, roughly, 10 ^ 18 guesses. Finally, if a, nonrandom, grammatical phrase is used, it should not be well known (e.g. book title, song title, lyric, famous quote, or a spelled out TLA)
It is good advice to suggest longer passwords. Unfortunately, many websites have a length limit which is too short. My credit union, a Silicon Valley financial institution no less, had an upper limit of 12 characters for passwords used for online banking. I discussed the problem with them about such a limit 3 months ago. As it happened they were in the midst of making changes directed towards making the site more secure. As a result, they increased the limit to 20 characters which is tolerable but not ideal.
A randomly generated password of 12 characters, using a good sized character set significantly larger than the set of alphanumeric characters, is fairly safe. This works well if you use a password manager and don't have to type, much less memorize, such a password. Passwords that don't require a lot of effort to memorize contain less entropy, so they need to be longer to be secure. I suggest 20 characters as a minimum. I used to use book titles along with a random 4 digit number. No longer is that safe. Now, for my 94 passwords, I use random sentence fragments from books along with numbers (I won't disclose any more, security through obscurity does have limited value). I keep all my passwords in an encrypted file. I find I can remember passwords that I use at least once a week. I don't need mobile access so I have not utilized a password manager. My solution is not the only good one and I do test it with a password cracker.
I believe the gist of this story is not so much blaming someone it's that both government and industry rely on an Internet which needs funding for critical areas involving security. One of the NSA's primary roles is ensuring the security of the Internet. That role can be filled by a government agency that provides, at least, funding to create and maintain the underlying code. There will never be trust in code used for confidentiality and authentication unless it is open source. Also, as many people are saying, the NSA can't play this role when it also has the role of spying on communications.
The KQED program about PSS quoted one LA sheriff as saying that this was a test and LASD is not planning to use PSS on a regular basis. Not that privacy was the issue that swayed them against adopting it, rather it was cost vs efficacy in solving crimes. The LASD does care about public opinion concerning surveillance insofar as public outcry can serve as an impediment to the adoption of any particular technology by the law enforcement.
Even if the cameras in the plane were higher resolution facial recognition still cannot be applied to the images. That is, unless you can get the person to look up at the plane somehow. One of the LA sheriffs noted that PSS posed the least intrusive surveillance technique compared to the other new technologies coming in to play.
Whenever there is a credit card or debit card payment involved, a retailer or any other sort of business has a contract with the credit card companies and the banks that issue such cards. That contract obligates the retail business to comply with the PCI-DSS (Payment Card Industry Digital Security Standard). When there is a security breach there will be an investigation which will determine whether the retailer was in full compliance with PCI-DSS. If not, they will be liable, rather than the banks, for losses due to resulting fraud. Also, additional fines can be levied against the retailer. Given this, it is a bit odd that the FTC is trying apply civil penalties via a lawsuit outside of this existing mechanism. There is a lot of argument about whether compliance with PCI-DSS is enough to prevent most attacks. Compliance is expensive, time consuming, and the bureaucratic line item approach misses out on some intuitively obvious ways to better ensure security. Yet, it is at least a fairly comprehensive standard which the FTC is lacking.
Although the prosecution was brought by federal prosecutors within the DOJ, this case was initially investigated by the FBI at the behest of AT&T. I don't think one can minimize the influence of AT&T in getting the government to pursue this case, although the details of that influence will probably never be known publicly. The case was, and is, such a weak one that never should have been pursued. Recall two people were charged; Andrew Auernheimer (Weev) and Daniel Spitler. Spitler pleaded guilty to the charges and was sentenced to 3 years probation on January 24, 2014. Compare that to 41 months of prison for Weev. This is yet another example of how people are severely punished, particularly in federal court, for fighting the charges against them.
I will provide the following timeline that shows how quickly the FBI got involved.
June 3, 2010 - June 8, 2010: Spitler and Weev collect email address/ICCID pairs.
June 6, 2010: Weev send emails to a handful of top media personnel whose emails were collected. He briefly explains how he came to know their email address and invites them to interview him. Weev explained that this was his way of, indirectly notifying AT&T of the security vulnerability.
June 7, 2010: AT&T is notified of the security breach by a “business customer” who is not identified by AT&T.
June 8, 2010: AT&T has stated that they fixed this vulnerability, by Tuesday, within hours of being notified of the problem. They did this by disabling or removing the code which pre-populated the log-in page with an email address.
June 9, 2010: Weev contacts Ryan Tate of Gawker gives him the list of email address/ICCID pairings and details about their uncovering of AT&T's security hole. Gawker publishes and article that very afternoon including a handful of redacted pairings that were for notable people.
June 10, 2010: Gawker is contacted by the FBI and issued a formal preservation of evidence notice.
You can see that the FBI was involved very early on. I can imagine that they were contacted by some executive at AT&T as soon as AT&T had learned of the breach.
I am not sure the NSA actually needs that site. It may be just convenient as there are already facilities there now. The original rationale for using the radio-quiet area is to avoid radio interference in detecting weak signals reflected off the moon (USSR, cold war, etc...). The original concept of a 600 ft. radio dish antenna was never completed. I will hazard a guess that it was replaced by the utilization of Diego Garcia which, via atmospheric properties and positioning on the globe, allowed for monitoring of Russian radio signals. So, Sugar Grove's main purpose for the NSA has been as a COMSAT receiving station. These are our own satellites so weak signal detection is not an issue. I don't think that the fact this NSA facility is in a quiet area is going to help them detect nefarious groups who are trying to snoop on the NSA. Most such radio frequency snooping is passive. Better solutions are encryption and physical protections like the false facade of the Puzzle Palace.
Oops..., Senator Joe Manchin says the upper part of the base will still be active after 2015. So, they still need to collect data reflected off the moon.
Electro-sensitives can celebrate as Sugar Grove is going to be shutdown by September 30, 2015. http://cryptome.org/dodi/2013/opnav-5400-2215.pdf Oh wait, the liability shift for adopting chip and pin credit cards happens on October 1st of that year. They are going to have to abandon their credit cards as well.
Firstly, I want to say that I agree, wholeheartedly, with the ACLU and these databases should not retain detain data for more than a few days. I will now argue as a devil's advocate and point out that there is a good reason to retain ALPR data for years. The value that law enforcement sees in long term data retention are for cases that are unknown at the time the ALPR record is made. For any new suspect the database may provide a number of records that may or may not be useful in the investigation. Even if the value of that data is minimal you cannot argue that it is a waste of resources for the simple reason that technology has made the recording of license plates, creation and retention of this data in a database so damn cheap.
The argument against long term retention should be focused on the fact that this is yet another method to automatically track the movements of nearly all citizens. Currently, the majority of people will have their cars recorded, either not at all or only a few times a year. That is not so worrisome. However, these use of these systems is rapidly proliferating and when your car is recorded everyday that data, whether in the hands of government or a private company is rather worrisome.
The cameras that are part of the ALPR system take photos that can include enough detail to recognize the driver, if there is one, and the photo is taken from the front of the vehicle. There is an option to store these photos as part of a record in the database. That recognition may not be automated facial recognition, but if a record is being investigated, for whatever reason, the human investigator has a picture of the driver available.
So, the charge is "disturbing the peace", and that is a vague catch-all for any so-called crime that doesn't violate some, more specific, law. The rationale for this charge can't be arbitrary though. If the same behavior was perfectly legal outside of school then you can't define it as a crime simply because it violates school rules. Otherwise, the school district is in a position of legislating, creating new crimes, and they do not have that authority. A criminal charge for violating school rules against toy guns should not stand, precisely for this reason. The other underlying behavior for the crime was intending to threaten a teacher. Actually threatening a teacher could be a real crime but there was no actual threat. The boys were arrested before a threat could take place and just intending to threaten is not a crime. This needs to be challenged in court and the DA needs to be taught a basic lesson about the justice system in America.
I would be very surprised if there was not a national ALPR database already. Most of the funds that police departments use to purchase these systems are grants from various agencies under DHS. The main focus of those grants is providing for protection of critical infrastructure. It would be hard to believe that such grants did not come without strings attached requiring the sharing of collected information. A large number of documents requesting and issuing these grants for various police departments across the country was requested by the ACLU for their report which was issued last year. There is always some redacted information in the contracts related to the grants. Already, a lot of information is shared with the fusion centers operated under DHS purview. Why not just one more step of aggregating ALPR database information from the 72 fusion centers. This proposal may be just to unify and improve the database so that there are no barriers in merging information from disparate databases and improving the ability to conduct queries.
I can tell you of an instance where AT&T was throttling my bandwidth and the evidence is fairly convincing. This happened a few years ago at my home. I have DSL service through AT&T but I use a different ISP, namely Cruzio who has a contract with AT&T for providing the DSL infrastructure and service to each customer. My service rate is nominally 1.5 Mbps but I generally test out at a download max of 1.3 Mbps. I live in a rural area so, yep, I can't get faster DSL service. One day I noticed videos were pausing unexpectedly. I ran a speed test. The max download rate was consistently 384 Kbps. That was a very suspicious number and suggested that my line had been capped at that rate at the central office. I called Cruzio and asked if they knew what the problem was. The service rep said that they had a number of Cruzio customers who had recently run into that exact problem in my area. They said they would call AT&T. Less than ten minutes later my max download rate was back up to 1.3 Mbps. This problem has not occurred again. My theory is that AT&T had a capacity problem at that central office as folk in the area were increasingly adopting DSL in place of dial-up internet service. AT&T decided to handle this in a surreptitious way by capping individual DSL rates at a fraction of what my agreement with Cruzio stated and, undoubtedly, in violation of Cruzio's contract with AT&T. Those who complained were uncapped but those who didn't suffered, perhaps unknowingly, with a lower rate. This may well have been temporary until equipment updates at the central office increased total throughput. It does show that the telecoms are willing to quietly shaft their customers. Mine was a general problem, not just throughput from a particular site like Netflix. In the latter case, as this article points out, it is impossible for the end-user to know if a rate problem is not due to congestion rather than deliberate capping.
This is actually a good idea as long as the records are anonymized well enough. Anonymization of medical data can be difficult when dealing with rare diseases or medical conditions. Let's take a look at how the care.data system handles this.
"Your date of birth, full postcode, NHS Number and gender rather than your name will be used to link your records in a secure system, managed by the HSCIC. Once this information has been linked, a new record will be created. This new record will not contain information that identifies you. The type of information shared, and how it is shared, is controlled by law and strict confidentiality rules."
See, the database will not contain information that identifies you. Problem solved.
My comment about the Google bot was rhetorical. Of course, it is not up to the web spiders, or the companies that build them, to try to figure out if the builders of the website really wanted a page to be public or not. Outside of the convention of robots.txt, if a bot can read a page then it gets read, indexed, and cached. If ANSES had done the authentication and authorization correctly they wouldn't even need to use robots.txt. My point was that it is equally absurd to penalize a person who reads and caches a webpage that has no effective protection against unauthorized persons reading it. There is a cultural assumption that pages on the Internet are for public consumption unless there is some technical method which prevents straightforward navigation and reading. This is contrary to the usual trespassing analogies where the cultural assumption is that a place is private property and you are trespassing unless you have explicit permission.
Here, we have a situation where attempted webpage protection was completely ineffective. This allowed Google, and any other bot or human, to read, index, and cache a large set of pages that were intended to be private. You can't punish someone for doing a search and then reading the resulting webpages that are unprotected. Laurelli is being punished because, after reading those pages, he travels back to the home page and sees that ANSES intended those pages to be accessible only after logging in. This is very screwed up justice and I will dare to offer this trespassing analogy:
Suppose you have a park in the US which seems to be public. You walk into the park, wander around, and then leave through the main entrance. At this entrance you turn around and there is a sign, in Russian, which says "no trespassing". Is the government only going to prosecute those trespassers who can speak Russian?
Effectively, there was no security, but why didn't the Google bot notice there was a log-in required on the home page? A person not as technically astute as Laurelli would not have known they weren't supposed to be looking at these documents. It seems that the security was supposed to be limiting access to URLs to only those who logged in on the home page. I am speculating that the mistake was that at the same time someone was logged in (who had the password "Fatalitas") the Google bot came by to index all the linked pages without needing a separate login. Once indexed, and also in the Google cache, this allowed any person access to the pages. Laurelli has been fined only because he admitted traveling back to the home page and noting there was an authentication step. It seems that knowledge is enough to warrant a penalty. This goes beyond the matter of criminal intent being a required element of a crime. What we have here is mind boggling in that a crime is only a crime if you know it is a crime.
"a cryptographic function known as a "hash" -- a transformation that converts it into a unique string of characters -- it produces an encrypted version of the sender's message, ready to be decrypted with the recipient's key."
When I read this my skepticism reached overload. A hash is a one-way mathematical function, and by definition, cannot be decrypted with a key. I figured that maybe it was just Andy Greenberg who misunderstood the algorithm here. That appears to be true, but I will cut him some slack because Bram Cohen's explanation of this on Github sucks, to be frank. From what I think I understand the algorithm to be, it is rather clever. It does go to show that sometimes smart people don't have the ability to explain well what they know.
I haven't read the code yet, only the textual notes. So, this may not be correct, but here goes. The elements are:
-cover text, for which there exists a set of short alternate segments. For each of these segments there is a single alternate which makes as much sense as the original.
-A shared cryptographic key.
-a value, which is the message to be hidden.
-SHA3 cryptographic hash algorithm
-a custom stream cipher which is a variation of AES in Output FeedBack mode (OFB). An important aspect of the algorithm for this stream cipher is that the set of segment alternates can be found, by the "encoding" portion of the program, that, with the chosen key, will produce the desired value (the message) as the first part of the encrypted output of the cipher.
The first step, for the sender, is to encrypt the cover text with chosen alternates using the shared key and an initialization vector (what Cohen is calling the salt), and AES in OFB mode. This initialization vector is created by using the first 4 bytes of the SHA3 hash of the chosen cover text.
There is a packing step which adds a length prefix and a checksum. The resulting data can be posted on a public website.
The message receiver will also have the shared key and can apply the custom stream cipher to reveal the message.
There are some details which I am still unclear about. I am not a cryptologist, so I cannot evaluate this scheme.
On the post: US Government Is Paying To Undermine Internet Security, Not To Fix It
Re:
The current state of password cracking allows secure use of passphrases though. A coarse attack against a passphrase using a dictionary of 20,000 words requires an effort of 20,000 ^ N, where N is the number of words. If N is 2 that effort is 400,000,000 (actually 200 million on average) This is still not secure. 3 words requires an average effort of 4 trillion guesses. This is still not secure particularly if the words are not random but a sentence fragment. The security can be increased with mangling but it is better to choose a basic length without considering mangling. 5 random words requires and average effort of 1.6 x 10 ^ 21. This is very roughly equivalent to a binary key of length 70 (70 bits of entropy) and with mangling can approach a password that, depending on the hashing algorithm employed, even the NSA will, currently, have a hard time cracking in a reasonable time. If one uses a larger dictionary, say 1 million words including all sorts of technical words, a random 3 word passphrase requires an effort of, roughly, 10 ^ 18 guesses.
Finally, if a, nonrandom, grammatical phrase is used, it should not be well known (e.g. book title, song title, lyric, famous quote, or a spelled out TLA)
On the post: US Government Is Paying To Undermine Internet Security, Not To Fix It
password length
A randomly generated password of 12 characters, using a good sized character set significantly larger than the set of alphanumeric characters, is fairly safe. This works well if you use a password manager and don't have to type, much less memorize, such a password. Passwords that don't require a lot of effort to memorize contain less entropy, so they need to be longer to be secure. I suggest 20 characters as a minimum. I used to use book titles along with a random 4 digit number. No longer is that safe. Now, for my 94 passwords, I use random sentence fragments from books along with numbers (I won't disclose any more, security through obscurity does have limited value). I keep all my passwords in an encrypted file. I find I can remember passwords that I use at least once a week. I don't need mobile access so I have not utilized a password manager. My solution is not the only good one and I do test it with a password cracker.
On the post: US Government Is Paying To Undermine Internet Security, Not To Fix It
Re:
On the post: LA Sheriff's Dept. On New Surveillance Program: We Knew The Public Wouldn't Like It, So We Kept It A Secret
Even if the cameras in the plane were higher resolution facial recognition still cannot be applied to the images. That is, unless you can get the person to look up at the plane somehow. One of the LA sheriffs noted that PSS posed the least intrusive surveillance technique compared to the other new technologies coming in to play.
On the post: Court Says FTC Can Go After Companies Who Get Hacked For Their Weak Security Practices
There is a lot of argument about whether compliance with PCI-DSS is enough to prevent most attacks. Compliance is expensive, time consuming, and the bureaucratic line item approach misses out on some intuitively obvious ways to better ensure security. Yet, it is at least a fairly comprehensive standard which the FTC is lacking.
On the post: Prosecutors Admit They Don't Understand What Weev Did, But They're Sure It's Like Blowing Up A Nuclear Plant
Re:
I will provide the following timeline that shows how quickly the FBI got involved.
June 3, 2010 - June 8, 2010: Spitler and Weev collect email address/ICCID pairs.
June 6, 2010: Weev send emails to a handful of top media personnel whose emails were collected. He briefly explains how he came to know their email address and invites them to interview him. Weev explained that this was his way of, indirectly notifying AT&T of the security vulnerability.
June 7, 2010: AT&T is notified of the security breach by a “business customer” who is not identified by AT&T.
June 8, 2010: AT&T has stated that they fixed this vulnerability, by Tuesday, within hours of being notified of the problem. They did this by disabling or removing the code which pre-populated the log-in page with an email address.
June 9, 2010: Weev contacts Ryan Tate of Gawker gives him the list of email address/ICCID pairings and details about their uncovering of AT&T's security hole. Gawker publishes and article that very afternoon including a handful of redacted pairings that were for notable people.
June 10, 2010: Gawker is contacted by the FBI and issued a formal preservation of evidence notice.
You can see that the FBI was involved very early on. I can imagine that they were contacted by some executive at AT&T as soon as AT&T had learned of the breach.
On the post: Town Built Around No WiFi/Radio Waves Rules Is Right Next To NSA Snoop Center
Re:
On the post: Town Built Around No WiFi/Radio Waves Rules Is Right Next To NSA Snoop Center
Re:
On the post: Town Built Around No WiFi/Radio Waves Rules Is Right Next To NSA Snoop Center
http://cryptome.org/dodi/2013/opnav-5400-2215.pdf
Oh wait, the liability shift for adopting chip and pin credit cards happens on October 1st of that year. They are going to have to abandon their credit cards as well.
On the post: No, Government Computers Won't Suddenly Be Vulnerable To Hackers On April 8; They Already Are
Re: Re: Re: Re: Re:
On the post: ACLU Battles Connecticut Law Enforcement Agencies Over Retention Of Licence Plate Reader Data
The argument against long term retention should be focused on the fact that this is yet another method to automatically track the movements of nearly all citizens. Currently, the majority of people will have their cars recorded, either not at all or only a few times a year. That is not so worrisome. However, these use of these systems is rapidly proliferating and when your car is recorded everyday that data, whether in the hands of government or a private company is rather worrisome.
On the post: ACLU Battles Connecticut Law Enforcement Agencies Over Retention Of Licence Plate Reader Data
Re: Re: Re:
On the post: Zero Tolerance Nets Two 11-Year Old Boys Juvenile Criminal Charges For Bringing A Toy Gun To School
What law was broken?
On the post: The DHS Sends Out The Call For A National License Plate Database
On the post: Netflix Rather Quietly Admits Verizon Isn't Throttling Netflix Streams
My theory is that AT&T had a capacity problem at that central office as folk in the area were increasingly adopting DSL in place of dial-up internet service. AT&T decided to handle this in a surreptitious way by capping individual DSL rates at a fraction of what my agreement with Cruzio stated and, undoubtedly, in violation of Cruzio's contract with AT&T. Those who complained were uncapped but those who didn't suffered, perhaps unknowingly, with a lower rate. This may well have been temporary until equipment updates at the central office increased total throughput. It does show that the telecoms are willing to quietly shaft their customers. Mine was a general problem, not just throughput from a particular site like Netflix. In the latter case, as this article points out, it is impossible for the end-user to know if a rate problem is not due to congestion rather than deliberate capping.
On the post: UK Police And Companies Will Have Access To Database Of All England's Medical Records
"Your date of birth, full postcode, NHS Number and gender rather than your name will be used to link your records in a secure system, managed by the HSCIC. Once this information has been linked, a new record will be created. This new record will not contain information that identifies you. The type of information shared, and how it is shared, is controlled by law and strict confidentiality rules."
See, the database will not contain information that identifies you. Problem solved.
On the post: Frenchman Fined For 'Theft' And 'Fraudulent Retention' For Finding Health Docs Via A Google Search
Re: Re: Re:
My point was that it is equally absurd to penalize a person who reads and caches a webpage that has no effective protection against unauthorized persons reading it. There is a cultural assumption that pages on the Internet are for public consumption unless there is some technical method which prevents straightforward navigation and reading. This is contrary to the usual trespassing analogies where the cultural assumption is that a place is private property and you are trespassing unless you have explicit permission.
Here, we have a situation where attempted webpage protection was completely ineffective. This allowed Google, and any other bot or human, to read, index, and cache a large set of pages that were intended to be private. You can't punish someone for doing a search and then reading the resulting webpages that are unprotected. Laurelli is being punished because, after reading those pages, he travels back to the home page and sees that ANSES intended those pages to be accessible only after logging in. This is very screwed up justice and I will dare to offer this trespassing analogy:
Suppose you have a park in the US which seems to be public. You walk into the park, wander around, and then leave through the main entrance. At this entrance you turn around and there is a sign, in Russian, which says "no trespassing". Is the government only going to prosecute those trespassers who can speak Russian?
On the post: Frenchman Fined For 'Theft' And 'Fraudulent Retention' For Finding Health Docs Via A Google Search
Re:
On the post: A New Twist On Steganography From The Creator Of BitTorrent
Explanation is unclear
When I read this my skepticism reached overload. A hash is a one-way mathematical function, and by definition, cannot be decrypted with a key. I figured that maybe it was just Andy Greenberg who misunderstood the algorithm here. That appears to be true, but I will cut him some slack because Bram Cohen's explanation of this on Github sucks, to be frank. From what I think I understand the algorithm to be, it is rather clever. It does go to show that sometimes smart people don't have the ability to explain well what they know.
I haven't read the code yet, only the textual notes. So, this may not be correct, but here goes.
The elements are:
-cover text, for which there exists a set of short alternate segments. For each of these segments there is a single alternate which makes as much sense as the original.
-A shared cryptographic key.
-a value, which is the message to be hidden.
-SHA3 cryptographic hash algorithm
-a custom stream cipher which is a variation of AES in Output FeedBack mode (OFB). An important aspect of the algorithm for this stream cipher is that the set of segment alternates can be found, by the "encoding" portion of the program, that, with the chosen key, will produce the desired value (the message) as the first part of the encrypted output of the cipher.
The first step, for the sender, is to encrypt the cover text with chosen alternates using the shared key and an initialization vector (what Cohen is calling the salt), and AES in OFB mode. This initialization vector is created by using the first 4 bytes of the SHA3 hash of the chosen cover text.
There is a packing step which adds a length prefix and a checksum. The resulting data can be posted on a public website.
The message receiver will also have the shared key and can apply the custom stream cipher to reveal the message.
There are some details which I am still unclear about. I am not a cryptologist, so I cannot evaluate this scheme.
On the post: New York Times Suffers Redaction Failure, Exposes Name Of NSA Agent And Targeted Network In Uploaded PDF
Re: Re: Re: Re: Re: Re:
http://tv.adobe.com/watch/acrobat-tips-and-tricks/remove-sensitive-information-using-redaction- tools/
Next >>