Hubs are always slower, because every link to a hub is half duplex.
Meaning 100 megabits, but cannot transmit while receiving or receive while transmitting.
If trying to transmit and receive at full speed simultaneously, they will max out at 50megabits (before counting IP overhead)
A computer plugged into a switch directly usually has 100 megabits (or better) full duplex.
They can receive and transmit at the full 100 megabits simultaneously.
Half duplex is basically extinct.
Hubs are basically extinct.
But the connection is much slower, even if only one machine is plugged into it.
You convert it into a Base-52 or Base-26 representation, so all keybits are represented by ordinary letters such as A-Z (you might expand it a little to include common punctuation marks), and underline different characters on the page that correspond to digits of your key..
Also, you can XOR your 4096 key, by a truly random 4096 bit value.
Destroy the original key, and print the random number and the XOR result in two different books at completely different places.
You can XOR it a few more times, and divide the key into 5 or 6 pieces.
Then go to the library with a pencil one night, and (covertly) store a copy of your key in various books.
Making plenty of redundant copies of course, and visiting multiple libraries (for fear that part of your key would be unavailable due to someone else having checked out one of your bits).
Bonus points if you don't keep the key at your house, but instead print it in redundant pieces on a few dozen different library books at various different libraries.
Even if someone finds it, they'll have no idea what they are looking at.
Have the card itself execute decryption of the symmetric key without revealing the private key to the PC, when it's read.
It will probably be cheaper than the uber-expensive specialized scanner+software from this vendor, you'll need to be able to scan the "cheap" paper key, anyways
And more secure in that the private RSA key is not subject to being stolen from PC RAM, or by modifying the decryption program on the PC to capture the key.
This could be construed as speech.
A pro-environmentalist statement on one's own private property.
The constitution also prohibits depriving a person of their property rights without due process.
One of those implicit rights is (presumably) to decorate it as you like.
It could also be within the practice of religion.
A possible element of many christian religions is respecting God's earth -- then the city is interfering with the free exercise thereof, then and there on one's own private yard, where they should be free from government persecution of any sort.
Well, we don't have the full story.
We probably will never have the whole story.
Only the public record (if there is any), the part of the story each side's lawyers want to present in court.
My understanding is Childs' use of "no service password recovery" was some sort of compromise instead of leaving startup-config blank on certain devices.
But when it comes to CPEs in a sufficiently hostile environment, there are a few reasons to consider security measures like that (in some unusual circumstances).
It is not essential to assume Childs was a "sick" admin. There are circumstances that could exist, where in fact leaving starup-config blank or disabling password recovery would be best practice.
The thing is, when you administer a government network such as the FibreWAN, that extends to multiple branches and potentially distant offices, there will be some highly sensitive traffic, and strong security is a must.
Strong security means an unauthorized person cannot intercept or modify traffic, even if that unintended person is knowledgeable, works for one of the offices, and has malicious intent.
A common method of implementing security is to use access lists, VPNs, passwords, and AAA servers.
An example of the type of compromise that could occur: some lower-level employee working in a branch office attempting to conduct espionage for a foreign intel agency could covertly reboot
the edge router one night, with a console cable plugged in, break into Monitor mode...
Dump a copy of the startup config to their laptop for later review.
Bring the router back up, and sneak away with all the keys to the AAA server,
for example the RADIUS shared key.
Once they possess the keys, they may be able to sniff the traffic, or spoof RADIUS response packets, in order to let them login with full access to the device.
They can also copy the hashes of all local passwords in the config out elsewhere,
send Type 5 passwords through a brute force cracker (given enough time, the hacker will eventually know all passwords on that CPE).
Once they have that much, covertly gaining access to another CPE using the [probably] shared AAA key is a likely outcome.
Until you've logged in and checked the system there's no way to know if there is a proper startup-config in NVRAM or the status of password recovery. If you reset the router without a config, you are boned.
True, but the problem was they assumed he did the most evil thing he could have done, and put him in jail over baseless assumptions.
This is like throwing an admin in jail under the assumption they deleted/boot/* or C:\windows\system32\, just because they weren't prompt in revealing the admin password, and you can't check before Ctrl+ALt+Del rebooting the server.
My money says the local console/aux ports were disabled in the running configuration... to keep the "local morons" out.
I doubt that, they were probably passworded though. Console port is almost always what must be used to assist in diagnosing a failure that makes the device inaccessible from the normal inband management backbone. On enterprise networks, console is commonly unpassworded, even, when the device is in a locked server room, anyways (enable password required to make any serious changes).
In an emergency, the admin can't be dicking around on a core device with additional reboots that produce further impact, and management re-enablement procedures, they are under the gun to find out what's wrong exactly, as quickly as possible, and remediate.
The only place console ports might typically be locked out would be at CPEs, where the device is not in a cage or secured area.
But that's because CPEs are cheap and generally just swapped out, quickly, with a device that gets a backup config loaded, if they fail.
Since childs was reportedly using dial-in modems for Out of band access, there's no way these ports could be disabled, in that case.
Perhaps they were just morons and didn't realize they could use provided passwords and log in via console port without power cycling anything?
Even if Childs was incompetent and paranoid, that doesn't suggest intentional sabotage he was charged with....
Probably because a large tower is conspicious from a long distance, and potentially has an effect on a lot of other people, because it obstructs their unimpeded view of the sky.
It's nothing confined to your yard, like your choice of landscaping materials are.
One end of the cable plugs into the RJ45 port on the special port on the router / engine / supervisor module with the blue "Console" label
The other end plugs into a serial port (or more likely nowadays) a USB-to-Serial adapter connected to a PC or Laptop.
(Or the RJ45 port on a serial concentrator also refferred to as, terminal server / console server / serial console switch)
In any case, typically a command line prompt is presented to the serial port at a baud rate of 9600 (that makes old 14.4 modems seem fast).
And no network connectivity at all should be required to access this.
It's really the port used for emergencies, essential maintenance, or on secure devices (such as firewalls, especially) that are sometimes designated by the admin to be managed out of band only.
He might have foregone AAA on some critical devices, since he was not distributing access to many people but keeping it solely to himself... or (rather) since he [was] the only person who had all the keys. The prosecution's theory would kind of fall apart, if he was using AAA on the network, and admins' could add additional router admins at any time...
Reportedly an initial issue was childs' use of no service password-recovery.
As a security compromise to his preference of leaving startup config blank on certain devices, for security reasons.
If they had suspected he did this on the core routers, then there's no way they could risk rebooting them, without a lot of acceptable downtime and one hell of a disaster recovery plan...
However, that was likely a one-sided few favoring the prosecution. If Childs' in fact did not do that (and never said he did) remove startup configs or 'no service password-recovery' on physically secured core equipment, then their fears are not his fault..
Childs may have only told them what he was able to think about to mention.. kind of tough to fill someone in when you don't know what exactly they don't know, what they need to know, etc, etc, and they are impatient / arrogant (as many manager types can act, esp. when they think they are not getting what they want).
Also, you can't exactly search through your own notes, and write usable notes with access details intended for someone else, while sitting in a jail cell.
In other words, by overreacting, grabbing him, and throwing him in jail, they probably made it more difficult, or even impossible for him to provide the very type of information they were wanting....
'Just because you give someone a password doesn't mean that person knows how to use it. Childs's security measures would have included access lists that blocked attempted logins from non-specified IP addresses or subnets. I
Don't use a non-specified IP address.
Or more specifically: graph a console cable, plug it into the device, and do what you need to do.
That an unskilled individual would not necessarily be able to easily use them does not mean Childs did anything wrong.
In fact, this is exactly how things should be --
in case the password is compromised, there should be additional layers of defense (IP access lists),
to prevent convert abuse of accidentally leaked passwords.
No one password should ever give anyone free reign over a critical network, without at least also having physical access or passing through a designated management point.
Why do you think it's necessary to attend the game? I can just as easily learn those facts by watching the game on tv
Then you would not be subject to any terms printed on a ticket, since you didn't use a ticket, assuming you didn't otherwise enter a contractual relationship with NFL that had some strange requirements.
Also, copyright law -- in particular, 17 USC 204 -- requires that copyright transfer agreements and exclusive licenses be signed by the author, so your ticket theory is wrong anyway.
204 only requires that a memorandum be signed to make the transfer of existing rights effective.
One party can still contract with a second in a manner that the second is obliged to transfer any copyrights to the first party and refrain from exercising any of said copyrights pending transfer.
Or the first party can simply include assignment of copyright in the terms. The rights are secured in advance, it is not a transfer of rights, because the creator of the work never owns the rights in the first place.
A common example of this is a contract work situation. If you pay me to write a piece of software for you, and the contract owns copyright, when I finish the product and give you the code, the copyright is automatically yours.
There's no "transfer" statement or anything of the sort that needs to be assigned.
A verbal agreement to that effect is sufficient, and you (the recipient under the contract), automatically get the copyrights afforded you by the contract.
You certified the agreement by taking the ticket and presenting it for admission.
The NFL received consideration (cash for the ticket), and you received consideration, access to the game.
The fact that an agent of you paid the fee is irrelevent.
The meeting of minds occured when you went there and presented the ticket to gain entry, and by seeking that they honor their end of the agreement, you confirm your own obligations.
I do not think Microsoft views Linux as a legitimate competitor.
Besides, why risk losing credibility in manipulating search results, when there's a far simpler strategy?
Give yourself free ads for common search terms
What happens during a football game is just a collection of facts. Collections of facts are not copyrightable in most circumstances.
No. But the moment you fix that collection of facts into a tangible form, your fixation of those facts becomes copyrightable, and your contract with the NFL (as included on the ticket) automatically confers the exclusive ownership to the NFL, the moment you fix or state your account of those facts.
contract law usually limits contracts to something reasonable
Not really. As long as the contract cannot be taken as unconscionable or unenforable, it can be as unreasonable as they desire it to be.
A mere test of reason alone won't make a contract term unenforceable.
It is perfectly reasonable for me to go home and tell my friends what happened at the game. It is reasonable that I discuss it at work the next day, where many people could overhear it. Where do you draw the line?
You going home and telling your friends doesn't cause a work fixed into tangible form.
So now I'm reprinting a play-by-play of the game. Using just facts, I create my own artistic work and publish it on a blog. Is that fair use?
It could be fair use of the copyright you were forced to transfer to NFL.
But you posting it might (or might not) be in breach of your admission contract.
Now I'm earning money from my blog. Now ESPN posts my account without asking me. Did they violate my copyright or the NFLs'? The answer is, mine.
If your ticket conferred copyright to written accounts of the events witnessed to the NFL, then they have the rights to post it.
And probably the rights to send DMCA takedowns against your blog hosting provider.... (unfortunately)
Well, the situation is not the same unless they're claiming copyright for someon else's exclusive work. It can happen, but I doubt it is nearly as common as accidentally stamping the wrong patent number on a product.
Copyright does give them the right to prohibit descriptions or accounts of the game.
It's part of the contract terms to buy a ticket, that NFL gets assigned those copyrights.
When NFL owns the copyright to your description or account, they are in a position to dictate the terms under which that account or description may be used.
Copyright also does allow them to prohibit lending the item, and other things that are 'performance of the work'. Unless your activity falls under the library exception, you are prohibited by that from setting up a rental store and lending the CDs without proper licensing..
The only way you can do those things is through fair use, which essentially means you admit to copyright infringement, according to copyright law you infringe: but you are protected from liability for that infringement by the fair use doctrine.
There's no such thing as a questionable copyright notice, since copyrights are automatic and don't have to be registered, it is almost always possible for a legitimate copyright claim to be made by the manufacturer (even if in reality, their only creation is cover art and their process of selecting public domain material to fit in the middle --- their choices of which public domain materials to include in an anthology or collection are in fact copyrightable).
That last bit about "Subject to approval"
is a loophole... all Activision has to do is reject the final product, always find something wrong with it to deny approval.
By the time the person's condition reaches the need for it, the dialysis itself might accelerate death, before the treatment begins to succeed..
But suppose you could chemically de-activate the antibiotic in the blood stream before it reaches each kidney, and re-activate it in the blood stream afterwards in a safe place...
Then I suppose, if the infection evolved intelligently, it would seek to infect the kidneys where it would be safer from the antibiotic.......
Just use traffic shaping, on the workstation itself
Hubs are always slower, because every link to a hub is half duplex. Meaning 100 megabits, but cannot transmit while receiving or receive while transmitting. If trying to transmit and receive at full speed simultaneously, they will max out at 50megabits (before counting IP overhead)
A computer plugged into a switch directly usually has 100 megabits (or better) full duplex. They can receive and transmit at the full 100 megabits simultaneously.
Half duplex is basically extinct. Hubs are basically extinct.
But the connection is much slower, even if only one machine is plugged into it.
You convert it into a Base-52 or Base-26 representation, so all keybits are represented by ordinary letters such as A-Z (you might expand it a little to include common punctuation marks), and underline different characters on the page that correspond to digits of your key..
Also, you can XOR your 4096 key, by a truly random 4096 bit value.
Destroy the original key, and print the random number and the XOR result in two different books at completely different places.
You can XOR it a few more times, and divide the key into 5 or 6 pieces.
Then go to the library with a pencil one night, and (covertly) store a copy of your key in various books.
Making plenty of redundant copies of course, and visiting multiple libraries (for fear that part of your key would be unavailable due to someone else having checked out one of your bits).
Bonus points if you don't keep the key at your house, but instead print it in redundant pieces on a few dozen different library books at various different libraries.
Even if someone finds it, they'll have no idea what they are looking at.
How about SmartCards and a smartcard reader?
Have the card itself execute decryption of the symmetric key without revealing the private key to the PC, when it's read.
It will probably be cheaper than the uber-expensive specialized scanner+software from this vendor, you'll need to be able to scan the "cheap" paper key, anyways
And more secure in that the private RSA key is not subject to being stolen from PC RAM, or by modifying the decryption program on the PC to capture the key.
This could be construed as speech. A pro-environmentalist statement on one's own private property.
The constitution also prohibits depriving a person of their property rights without due process. One of those implicit rights is (presumably) to decorate it as you like.
It could also be within the practice of religion. A possible element of many christian religions is respecting God's earth -- then the city is interfering with the free exercise thereof, then and there on one's own private yard, where they should be free from government persecution of any sort.
Well, we don't have the full story. We probably will never have the whole story. Only the public record (if there is any), the part of the story each side's lawyers want to present in court.
My understanding is Childs' use of "no service password recovery" was some sort of compromise instead of leaving startup-config blank on certain devices.
But when it comes to CPEs in a sufficiently hostile environment, there are a few reasons to consider security measures like that (in some unusual circumstances).
It is not essential to assume Childs was a "sick" admin. There are circumstances that could exist, where in fact leaving starup-config blank or disabling password recovery would be best practice.
The thing is, when you administer a government network such as the FibreWAN, that extends to multiple branches and potentially distant offices, there will be some highly sensitive traffic, and strong security is a must.
Strong security means an unauthorized person cannot intercept or modify traffic, even if that unintended person is knowledgeable, works for one of the offices, and has malicious intent.
A common method of implementing security is to use access lists, VPNs, passwords, and AAA servers.
An example of the type of compromise that could occur: some lower-level employee working in a branch office attempting to conduct espionage for a foreign intel agency could covertly reboot the edge router one night, with a console cable plugged in, break into Monitor mode...
Dump a copy of the startup config to their laptop for later review. Bring the router back up, and sneak away with all the keys to the AAA server, for example the RADIUS shared key.
Once they possess the keys, they may be able to sniff the traffic, or spoof RADIUS response packets, in order to let them login with full access to the device.
They can also copy the hashes of all local passwords in the config out elsewhere, send Type 5 passwords through a brute force cracker (given enough time, the hacker will eventually know all passwords on that CPE).
Once they have that much, covertly gaining access to another CPE using the [probably] shared AAA key is a likely outcome.
Until you've logged in and checked the system there's no way to know if there is a proper startup-config in NVRAM or the status of password recovery. If you reset the router without a config, you are boned.
True, but the problem was they assumed he did the most evil thing he could have done, and put him in jail over baseless assumptions. This is like throwing an admin in jail under the assumption they deleted /boot/* or C:\windows\system32\, just because they weren't prompt in revealing the admin password, and you can't check before Ctrl+ALt+Del rebooting the server.
My money says the local console/aux ports were disabled in the running configuration... to keep the "local morons" out.
I doubt that, they were probably passworded though. Console port is almost always what must be used to assist in diagnosing a failure that makes the device inaccessible from the normal inband management backbone. On enterprise networks, console is commonly unpassworded, even, when the device is in a locked server room, anyways (enable password required to make any serious changes).
In an emergency, the admin can't be dicking around on a core device with additional reboots that produce further impact, and management re-enablement procedures, they are under the gun to find out what's wrong exactly, as quickly as possible, and remediate.
The only place console ports might typically be locked out would be at CPEs, where the device is not in a cage or secured area. But that's because CPEs are cheap and generally just swapped out, quickly, with a device that gets a backup config loaded, if they fail.
Since childs was reportedly using dial-in modems for Out of band access, there's no way these ports could be disabled, in that case.
Perhaps they were just morons and didn't realize they could use provided passwords and log in via console port without power cycling anything?
Even if Childs was incompetent and paranoid, that doesn't suggest intentional sabotage he was charged with....
I said grab a console cable and plug it into the router, to log in. I didn't say anything about "grab a console cable and reboot the router"
One day some abusive admin did a rm -rf / or deltree C:\
Those bastards... that's why root should have no powers.
And the 'rm' command needs to be redacted.
Probably because a large tower is conspicious from a long distance, and potentially has an effect on a lot of other people, because it obstructs their unimpeded view of the sky.
It's nothing confined to your yard, like your choice of landscaping materials are.
What is its function?
I meant grab a console cable :)
One end of the cable plugs into the RJ45 port on the special port on the router / engine / supervisor module with the blue "Console" label
The other end plugs into a serial port (or more likely nowadays) a USB-to-Serial adapter connected to a PC or Laptop.
(Or the RJ45 port on a serial concentrator also refferred to as, terminal server / console server / serial console switch)
In any case, typically a command line prompt is presented to the serial port at a baud rate of 9600 (that makes old 14.4 modems seem fast).
And no network connectivity at all should be required to access this.
It's really the port used for emergencies, essential maintenance, or on secure devices (such as firewalls, especially) that are sometimes designated by the admin to be managed out of band only.
He might have foregone AAA on some critical devices, since he was not distributing access to many people but keeping it solely to himself... or (rather) since he [was] the only person who had all the keys. The prosecution's theory would kind of fall apart, if he was using AAA on the network, and admins' could add additional router admins at any time...
Reportedly an initial issue was childs' use of no service password-recovery. As a security compromise to his preference of leaving startup config blank on certain devices, for security reasons.
If they had suspected he did this on the core routers, then there's no way they could risk rebooting them, without a lot of acceptable downtime and one hell of a disaster recovery plan...
However, that was likely a one-sided few favoring the prosecution. If Childs' in fact did not do that (and never said he did) remove startup configs or 'no service password-recovery' on physically secured core equipment, then their fears are not his fault..
Childs may have only told them what he was able to think about to mention.. kind of tough to fill someone in when you don't know what exactly they don't know, what they need to know, etc, etc, and they are impatient / arrogant (as many manager types can act, esp. when they think they are not getting what they want).
Also, you can't exactly search through your own notes, and write usable notes with access details intended for someone else, while sitting in a jail cell.
In other words, by overreacting, grabbing him, and throwing him in jail, they probably made it more difficult, or even impossible for him to provide the very type of information they were wanting....
'Just because you give someone a password doesn't mean that person knows how to use it. Childs's security measures would have included access lists that blocked attempted logins from non-specified IP addresses or subnets. I
Don't use a non-specified IP address.
Or more specifically: graph a console cable, plug it into the device, and do what you need to do.
That an unskilled individual would not necessarily be able to easily use them does not mean Childs did anything wrong.
In fact, this is exactly how things should be -- in case the password is compromised, there should be additional layers of defense (IP access lists), to prevent convert abuse of accidentally leaked passwords.
No one password should ever give anyone free reign over a critical network, without at least also having physical access or passing through a designated management point.
Why do you think it's necessary to attend the game? I can just as easily learn those facts by watching the game on tv
Then you would not be subject to any terms printed on a ticket, since you didn't use a ticket, assuming you didn't otherwise enter a contractual relationship with NFL that had some strange requirements.
Also, copyright law -- in particular, 17 USC 204 -- requires that copyright transfer agreements and exclusive licenses be signed by the author, so your ticket theory is wrong anyway.
204 only requires that a memorandum be signed to make the transfer of existing rights effective.
One party can still contract with a second in a manner that the second is obliged to transfer any copyrights to the first party and refrain from exercising any of said copyrights pending transfer.
Or the first party can simply include assignment of copyright in the terms. The rights are secured in advance, it is not a transfer of rights, because the creator of the work never owns the rights in the first place.
A common example of this is a contract work situation. If you pay me to write a piece of software for you, and the contract owns copyright, when I finish the product and give you the code, the copyright is automatically yours.
There's no "transfer" statement or anything of the sort that needs to be assigned. A verbal agreement to that effect is sufficient, and you (the recipient under the contract), automatically get the copyrights afforded you by the contract.
You certified the agreement by taking the ticket and presenting it for admission.
The NFL received consideration (cash for the ticket), and you received consideration, access to the game. The fact that an agent of you paid the fee is irrelevent.
The meeting of minds occured when you went there and presented the ticket to gain entry, and by seeking that they honor their end of the agreement, you confirm your own obligations.
I do not think Microsoft views Linux as a legitimate competitor.
Besides, why risk losing credibility in manipulating search results, when there's a far simpler strategy? Give yourself free ads for common search terms
Example: content management, sql, mail server, antivirus, email account, instant messenger
What happens during a football game is just a collection of facts. Collections of facts are not copyrightable in most circumstances.
No. But the moment you fix that collection of facts into a tangible form, your fixation of those facts becomes copyrightable, and your contract with the NFL (as included on the ticket) automatically confers the exclusive ownership to the NFL, the moment you fix or state your account of those facts.
contract law usually limits contracts to something reasonable
Not really. As long as the contract cannot be taken as unconscionable or unenforable, it can be as unreasonable as they desire it to be.
A mere test of reason alone won't make a contract term unenforceable.
It is perfectly reasonable for me to go home and tell my friends what happened at the game. It is reasonable that I discuss it at work the next day, where many people could overhear it. Where do you draw the line?
You going home and telling your friends doesn't cause a work fixed into tangible form.
So now I'm reprinting a play-by-play of the game. Using just facts, I create my own artistic work and publish it on a blog. Is that fair use?
It could be fair use of the copyright you were forced to transfer to NFL. But you posting it might (or might not) be in breach of your admission contract.
Now I'm earning money from my blog. Now ESPN posts my account without asking me. Did they violate my copyright or the NFLs'? The answer is, mine.
If your ticket conferred copyright to written accounts of the events witnessed to the NFL, then they have the rights to post it. And probably the rights to send DMCA takedowns against your blog hosting provider.... (unfortunately)
In Microsoft's case, the only tool they have is a nuclear weapon.
Well, the situation is not the same unless they're claiming copyright for someon else's exclusive work. It can happen, but I doubt it is nearly as common as accidentally stamping the wrong patent number on a product.
Copyright does give them the right to prohibit descriptions or accounts of the game. It's part of the contract terms to buy a ticket, that NFL gets assigned those copyrights. When NFL owns the copyright to your description or account, they are in a position to dictate the terms under which that account or description may be used.
Copyright also does allow them to prohibit lending the item, and other things that are 'performance of the work'. Unless your activity falls under the library exception, you are prohibited by that from setting up a rental store and lending the CDs without proper licensing..
The only way you can do those things is through fair use, which essentially means you admit to copyright infringement, according to copyright law you infringe: but you are protected from liability for that infringement by the fair use doctrine.
There's no such thing as a questionable copyright notice, since copyrights are automatic and don't have to be registered, it is almost always possible for a legitimate copyright claim to be made by the manufacturer (even if in reality, their only creation is cover art and their process of selecting public domain material to fit in the middle --- their choices of which public domain materials to include in an anthology or collection are in fact copyrightable).
That last bit about "Subject to approval" is a loophole... all Activision has to do is reject the final product, always find something wrong with it to deny approval.
Sort of like gets done with iPhn Appstor
By the time the person's condition reaches the need for it, the dialysis itself might accelerate death, before the treatment begins to succeed..
But suppose you could chemically de-activate the antibiotic in the blood stream before it reaches each kidney, and re-activate it in the blood stream afterwards in a safe place...
Then I suppose, if the infection evolved intelligently, it would seek to infect the kidneys where it would be safer from the antibiotic.......
The keyword is being developed
How many people will die of infection before the FDA gives them their seal of approval?
That's great, until you need to get a decent baud rate out of the port such as 76800 115200 or 230400 with proper behavior against finicky devices.
Or to be able to send a proper BREAK signal over your 'USB serial port' to get the router into rommon mode for password recovery.
Then suddenly (maybe) you find the $15 adapter isn't that good, and you need a feature that requires the $45 adapter that's hard-to-find .