Gregg's Nearly Interminable Adventures in Linux Land

[DMcCunney]

Originally Posted by Dngrsone

My goodness... I don't even remember. Let's see, Vulcan was a basic OS; the Machine itself was 24-bit mainframe with I think 512KB RAM. We used two HP7906 or 7907 hard drives (the 7906 had 18" platters, one fixed, one removable, total storage, a whopping 20MB; the 7907 went down to 8" platters, same capacity, IIRC). I think we calculated teh clock speed at something like 3MHz...

Job Control Language (JCL) ran on top of Vulcan, though we spent most of our time in Test Executive, which ran in the JCL shell.

The first machine I dealt with, back in the 80's, was an IBM 370 compatible. It ran OS/VS1, later converted to OS/MVS. The original machine had 2MB RAM and 16MB virtual memory, and supported about 500 remote 3270 terminals under CICS, when it worked. The area of the bank I worked in had brought in a complete IBM mainframe data center for just under $1 million, but to do so they went third-party and plug compatible, and because it wasn't genuine IBM kit, they couldn't run the latest IBM software. Outages were frequent.

I had a cartoon from Datamation in my cube, with a field engineer walking into a site asking "System been down long?", and addressing the question to a skeleton in a chair coated in cobwebs. Some made a copy of it, and put it up on the inside of the door to the VP of IT's office. He was not amused. He came up through the ranks, having started as a COBOL programmer. There was still a COBOL module on the system he maintained to keep his hand in. He sat down at the terminal do do a little work on "his" program, and Lo! The system was down. The systems programmer turned purple and sputtered when I told him about it.

Midway through my tenure, they upgraded to two IBM 4341s loosely coupled under JES2, and reliability soared. Around the time I left, the bank decided to centralize everything back at Division level and close the Region's data center. The Region had built out its own capacity in the first place to get out from under the two year backlog at Division, and it was hitting its stride and doing good stuff, but corporate had different ideas...

I was amused by IBM JCL. All of eight statements in the language, but it was a black art, and everyone used someone else's canned procs instead of trying to write their own. I got yelled at at one point because I tweaked the JCL on a job to boost it's prioriy, and got a "Don't do that!" reprimand from the VP of Applications Development. I'm not sure whether he was more upset that I'd done it, or that I wasn't a member of the IT staff. (I actually worked for Finance, and was their interface with IT.)

Access was via IBM's TSO. The original "GUI" was a third party product called ACEP, intended to be a substitute for IBM's SPF. When the bank upgraded to "real" SPF, I thought it a step backward.

I've got a crappy Motorola NVG510 that AT&T insists I use for U-verse. The built-in firewall is basic. I have an old gaming server with Smoothwall on it: CLAMAV and content filtering, usage restrictions and monitoring, etc... I can even set up a separate network for a server, if the son decides he wants to host a Minecraft server.

Problem is, I can't get the NVG510 to go into true bridge mode for my external firewall to work; hence the hacking part.

This might help: https://forums.att.com/t5/Third-Part...y/td-p/3434307

It's either that or try and get AT&T to admit that there are other (perhaps even better?) modems that will work on their lines.

There probably are. The question is whether AT&T would notice if you substituted one and what they would do if you did.
______
Dennis

[Dngrsone]

Originally Posted by DMcCunney

... I was amused by IBM JCL. All of eight statements in the language, but it was a black art, and everyone used someone else's canned procs instead of trying to write their own. I got yelled at at one point because I tweaked the JCL on a job to boost it's prioriy, and got a "Don't do that!" reprimand from the VP of Applications Development. I'm not sure whether he was more upset that I'd done it, or that I wasn't a member of the IT staff...

This might help: https://forums.att.com/t5/Third-Part...y/td-p/3434307

It was pretty simple, and considering the nature of our setup, it was kind of weird to have all this accounting and user control stuff for a mini that had only one terminal attached to it.

... well, actually, one of my coworkers had run wires out from his benches and set up his personal machine to act as a second terminal so he could monitor both benches from a back room, but back then (late '80s) that was akin to high sorcery. I know how he did it now, but I wasn't anything remotely network-savvy at the time, and yet I was one of only a few people who could write subroutines in JCL or (gasp) write a machine-language routine to be thumbed into the computer by hand.

That's an interesting link, and may just work, though I am leery about double-NATting. Then again, that's essentially what VPN and TOR is, right?

... I could try it the "legal" way before going the questionable route. Hurts nothing and no one (well, except for the inevitable hue and cry that the internet is down until I finish the job).

Thanks for the link!

[DMcCunney]

Originally Posted by Dngrsone

It was pretty simple, and considering the nature of our setup, it was kind of weird to have all this accounting and user control stuff for a mini that had only one terminal attached to it.

I have to assume the system could do enough to merit some form of job control, and that jobs might be submitted by different users, so some way of determining whose job it was was required.

I was amused by the way various assumptions were embedded in the IBM system. Earlier machines got fed jobs on punch cards, and you put a deck of cards into a card reader to load into the machine. Punch cards were largely gone when I got involved (though the data center did have a card reader). Instead, what was sent to the mainframe was a file of 80 column card images. I dealt with text files that were members of a partitioned data set, and created edited those files to make and submit jobs. The first cards in the deck were JCL statements to let the machine know it was a job, what its name was, what data sets it used, and what programs would be run as part of the job stream.

I also got more acquaintance than I wanted with error messages, and my experience was consistent. I'd get an error message, pull down the manual for a part of the system I was working with, go the the chapter governing the stuff I was using, and where I might expect an explanation of the error I encountered I'd find a reference to another manual I didn't ahve, no matter how many manuals I accumulated. (I never saw a complete set of manuals for an IBM 370 system.) VSAM errors were particular peeves. Next stop for me was Unix, and a complete set of manuals was three small sized binders occupying about a foot of space on the shelf. It was a revelation.

I later spent time in market research, and computing there began on mainframes. Data for market research projects was stored in IBM card/column format, and folks writing scripts in the specialized language implemented by the software had to plan where data was stored by card and columns. I did later encounter MR software that used an actual database as storage and could view it as you liked, but it wasn't the standard approach.

... well, actually, one of my coworkers had run wires out from his benches and set up his personal machine to act as a second terminal so he could monitor both benches from a back room, but back then (late '80s) that was akin to high sorcery. I know how he did it now, but I wasn't anything remotely network-savvy at the time, and yet I was one of only a few people who could write subroutines in JCL or (gasp) write a machine-language routine to be thumbed into the computer by hand.

An old friend once worked in a shop with two DEC PDP 1s. One was used for software development, and the other was used to play Spacewar. Someone decided only authorized users should be allowed to play Spacewar. So a programmer password protected the game, and told no one what it was. To play Spacewar, you had to go into the machine room, toggle front panel switches to put the system into single step debug mode, run Spacewar, step through the sequence till you got to where the password was input, figure out form what you saw what the password had to be, and enter it and play. If you were able to do that you were considered someone authorized to play Spacewar.

That's an interesting link, and may just work, though I am leery about double-NATting. Then again, that's essentially what VPN and TOR is, right?

Not really. I did something like that here previously.

What you essentially want to do is have the Motorola your access to the Uverse network, but have everything else done by a different system. So you put the Uverse into bridge mode, and all it does is serve as the gateway. Proving local IP addresses via DHCP, firewalling and the lake are done by another device, which gets a raw feed from the Motorala and sends stuff back out through it. You're taking the Motorola out of the loop as router, and having something else do it.

At one point, I was a Palm PDA connecting to my network. Everything else connected through WPA2 encryption, but the Palm only did WEP. I wasn't about to lower the security on my network, so I set up a second router configured as a bridge. It was seen as a trusted client by the main one. The PDA connected to the second modem, which forwarded the traffic through the master. DHCP, firewalling, and the like all happened on the main unit. When the PDA wasn't connected, the second unit was off. (I also turned on MAC address filtering on the second router so only the PDA could connect, and turned off SSID broadcast, to reduce likelihood of anyone in range seeing and connecting through it. I saw advice back when that if your security choices were WEP or none, use none so you wouldn't be under the illusion you were protected. )

... I could try it the "legal" way before going the questionable route. Hurts nothing and no one (well, except for the inevitable hue and cry that the internet is down until I finish the job).

The legal way I can see is used by the link I pointed to - put the Moto into bridge mode and have its functions performed by something else. I'd also look at getting a replacement modem. I'm not aware of anything special about Uverse that only the Motorola should work. This strikes me as AT&T being lazy and trying to reduce support issues. The question, if you found a compatible replacement, was whether they would even notice if you substituted.

I have Tor here, though I'm not playing with VPNs at the moment. Tor establishes an encrypted connection on a high port number to a Tor entry point, traffic is routed through multiple internal Tor network nodes, and proceeds to your destination from a Tor exit node. What the other ensd sees as your origin is the address of the Exit node, not yours. And your route through the internal Tor network changes periodically The intent is to make it impossible to trace your traffic back to you. In essence, Tor is an anonymous proxy with additional obfuscation in the proxy.

Thanks for the link!

You're welcome. Let us know how you make out.
______
Dennis

[Cinisajoy]

Dennis,
Uverse is supposed to be fiber optics as opposed to DSL and also it has a VOIP connection so your phone and Internet are one in the same.
Note: I have U verse but not a VOIP line.

[DMcCunney]

Originally Posted by Cinisajoy

Dennis,
Uverse is supposed to be fiber optics as opposed to DSL and also it has a VOIP connection so your phone and Internet are one in the same.
Note: I have U verse but not a VOIP line.

Doesn't matter. The question is whether you can change what is used to connect to their network. What the network itself is is a detail.

My first broadband, a dozen years or so ago, was actual DSL through what is now Verizon. Cable modem wasn't available where I was at the time, so DSL was the option. I installed a splitter jack on my copper phone line to connect a line to the vendor supplied DSL modem, and my phone line did both voice and data.

My cable provider is TimeWarner, and they recalled I'd enquired about a cable modem. When service became available where I was, they sent me a note. I could get a self-install kit with modem for $99, and pick it up across the street from my then office. Sold! I bought, connected, and was online in 15 minutes, at four times the speed of my DSL line for the same price. I kept the DSL line for a while as a high speed backup, and for a while I had two ISPs and two network interfaces. This mightily confused the software firewall I was running and required me to change firewalls. Cable service was reliable enough that I eventually dropped the DSL line.

TimeWarner was pusing VOIP at me as part of a triple play cable/Internet/VOIP bundle. I held off because it was a cost savings only if you made LD calls, and for practical purposes I didn't. Verizon's basic local loop charges eventually edged up to just over what VOIP would cost, and I switched. Immediately after, I found myself in a project that required living on the phone long distance to points south for a month. VOIP made that feasible.

TimeWarner is fending off Verizon FOIS and I benefit. My bandwidth has steadily increased at no change in my costs. Most recently, it was an upgrade for 20mbit "Turbo" service to 100mbit. That required a new combo modem/router that TimeWarner supplied, but I could use my own. They published a list of compatible devices users could install. (And they wanted you to, as one less thing they maintained.)

FIOS service isn't available where I am, but TimeWarner doesn't know that. And I'm happy with their service in any case. When I dropped the landline, I was delighted to say goodbye to Verizon, and have no desire to renew the relationship.

And I couldn't get a copper POTS line now in any case. VZ wants to make copper go away, and I don't blame them. Folks whose service was trashed by hurricane Sandy will not have it repaired. Their options are cell phone or VIOP via FIOS. Existing working copper will be maintained. New copper will not be run.
______
Dennis

[Dngrsone]

Originally Posted by DMcCunney

I have to assume the system could do enough to merit some form of job control, and that jobs might be submitted by different users, so some way of determining whose job it was was required.

Not really. I mean, as I said, one terminal. The computer was being used to operate an array of test equipment. There were a handful of subroutines that had to be running in the background (we're talking tens of individual components ranging from power supplies, measurement devices and switching to link it all together, using a half-dozen communications busses), but in the end, thre was only one user on the computer at any particular time.

I'm thinking that Harris found an excuse to offload some excess inventory and still fulfill their contract.

Originally Posted by DMcCunney

I was amused by the way various assumptions were embedded in the IBMsystem. Earlier machines got fed jobs on punch cards, and you put a deck of cards into a card reader to load into the machine. Punch cards were largely gone when I got involved (though the data cdenter did hve a card reader). Instead, what was sent to the mainframe was a file of 80 column card images. I dealt with text files that were members of a partitioned data set, and created edited those files to make and submit jobs. The first cards in the deck were JCL statements to let the machine know it was a job, what its name was, what data sets it used, and what programs would be run as part of the job stream.

Same here: there was a card reader interface, and theoretically there were card readers available (actually in one facility there was one on-hand, just not in my shop), and our software programs were written and structured for storage on punched cards, but really...

One of my instructors for the bench told us that our job (the one we were studying for) was a Cold war relic: should there be an nuclear blast (and accompanying EMP) that wiped out the data on our removable discs (those aforementioned 18" platters), then they could send out a card reader and stacks of cards and we could recompile all our programs and get running again. It sounded logical... in a Hollywood kind of way. Realistically, though, if the EMP were enough to penetrate several feet of steel to get to our platters and wipe them, then all the electronics in the entire test station would be crap, so they'd be shipping an entire bench or two, plus the software discs take up a lot less room and weigh less than an equivalent amount of punched-card data.

More likely is that Harris sold the card reader module with the computer or sold the government on the necessity to get rid of some of that excess inventory.

Originally Posted by DMcCunney

An old friend once worked in a shop with two DEC PDP [b]1[/i]s. One was used for software development, and the other was used to play Spacewar. Someone decided only authorized users should be allowed to play Spacewar. So a programmer password protected the game, and told no one what it was. To play Spacewar, you had to go into the machine room, toggle front panel switches to put the system into single step debug mode, run Spacewar, step through the sequence till you got to where the password was input, figure out form what you saw what the password had to be, and enter it and play. If you were able to do that you were considered someone authorized to play Spacewar.

Hahaha. Yeah, I learned a lot more than most because when one of our guys left the shop he slipped a little something into the benches to mess with us: whenever someone loaded up a specific program (for the easiest thing we had to work on), the bench would essentially tell the tech to quick cherry-picking and go work on something important.

It took a little research, but I finally figured out and found the script he wrote. It resided on the system disc and since the system disc would be searched before the removable drive, his script would execute first. The script was basically print message, go to removable drive and execute.

It's the same as substituting a custom script for an actual bash command by placing your script in a higher-priority directory.

Originally Posted by DMcCunney

What you essentially want to do is have the Motorola your access to the Uverse network, but have everything else done by a different system. So you put the Uverse into bridge mode, and all it does is serve as the gateway. Proving local IP addresses via DHCP, firewalling and the lake are done by another device, which gets a raw feed from the Motorala and sends stuff back out through it. You're taking the Motorola out of the loop as router, and having something else do it.

Exactly. The hack is basically exploiting a vulnerability on the modem and running commands as root, then using its nsh console to set it to true bridge mode, which is not possible through the public shell, by redirecting the PPP stream through one of the ethernet ports. reference

Originally Posted by DMcCunney

At one point, I was a Palm PDA connecting to my network. Everything else connected through WPA2 encryption, but the Palm only did WEP. I wasn't about to lower the security on my network, so I set up a second router configured as a bridge. It was seen as a trusted client by the main one. The PDA connected to the second modem, which forwarded the traffic through the master. DHCP, firewalling, and the like all happened on the main unit. When the PDA wasn't connected, the second unit was off. (I also turned on MAC address filtering on the second router so only the PDA could connect, and turned off SSID broadcast, to reduce likelihood of anyone in range seeing and connecting through it. I saw advice back when that if your security choices were WEP or none, use none so you wouldn't be under the illusion you were protected. )

Been there, done that: Palm T|X for close to eight years before I was forced to use a 'smart' phone.

Originally Posted by DMcCunney

The legal way I can see is used by the link I pointed to - put the Moto into bridge mode and have its functions performed by something else. I'd also look at getting a replacement modem. I'm not aware of anything special about Uverse that only the Motorola should work. This strikes me as AT&T being lazy and trying to reduce support issues. The question, if you found a compatible replacement, was whether they would even notice if you substituted.

I doubt they'd care, unless it caused them problems on their end. Of course, any issues I'd encounter would be pawned off on the unsupported modem and therefore be not their problem.

Originally Posted by DMcCunney

I have Tor here, though I'm not playing with VPNs at the moment. Tor establishes an encrypted connection on a high port number to a Tor entry point, traffic is routed through multiple internal Tor network nodes, and proceeds to your destination from a Tor exit node. What the other ensd sees as your origin is the address of the Exit node, not yours. And your route through the internal Tor network changes periodically The intent is to make it impossible to trace your traffic back to you. In essence, Tor is an anonymous proxy with additional obfuscation in the proxy.

Right. The onion routers remember where your machine is and route packets accordingly (I'm being overly simplistic, I know). Is this not the same as network address translation? I mean, my router remembers where my machine is at and routes the packets that belong to it accordingly... as far as AT&T knows (in my case), all data is being requested by and sent to one single IP, the router. It's the router's job to figure out which packets belong to my machine and which belong to the Playstation, or to the television.

Originally Posted by DMcCunney

Let us know how you make out.

I certainly will do so.

[Cinisajoy]

Originally Posted by Dngrsone

Not really. I mean, as I said, one terminal. The computer was being used to operate an array of test equipment. There were a handful of subroutines that had to be running in the background (we're talking tens of individual components ranging from power supplies, measurement devices and switching to link it all together, using a half-dozen communications busses), but in the end, thre was only one user on the computer at any particular time.

I'm thinking that Harris found an excuse to offload some excess inventory and still fulfill their contract.



Same here: there was a card reader interface, and theoretically there were card readers available (actually in one facility there was one on-hand, just not in my shop), and our software programs were written and structured for storage on punched cards, but really...

One of my instructors for the bench told us that our job (the one we were studying for) was a Cold war relic: should there be an nuclear blast (and accompanying EMP) that wiped out the data on our removable discs (those aforementioned 18" platters), then they could send out a card reader and stacks of cards and we could recompile all our programs and get running again. It sounded logical... in a Hollywood kind of way. Realistically, though, if the EMP were enough to penetrate several feet of steel to get to our platters and wipe them, then all the electronics in the entire test station would be crap, so they'd be shipping an entire bench or two, plus the software discs take up a lot less room and weigh less than an equivalent amount of punched-card data.

More likely is that Harris sold the card reader module with the computer or sold the government on the necessity to get rid of some of that excess inventory.




Hahaha. Yeah, I learned a lot more than most because when one of our guys left the shop he slipped a little something into the benches to mess with us: whenever someone loaded up a specific program (for the easiest thing we had to work on), the bench would essentially tell the tech to quick cherry-picking and go work on something important.

It took a little research, but I finally figured out and found the script he wrote. It resided on the system disc and since the system disc would be searched before the removable drive, his script would execute first. The script was basically print message, go to removable drive and execute.

It's the same as substituting a custom script for an actual bash command by placing your script in a higher-priority directory.



Exactly. The hack is basically exploiting a vulnerability on the modem and running commands as root, then using its nsh console to set it to true bridge mode, which is not possible through the public shell, by redirecting the PPP stream through one of the ethernet ports. reference



Been there, done that: Palm T|X for close to eight years before I was forced to use a 'smart' phone.




I doubt they'd care, unless it caused them problems on their end. Of course, any issues I'd encounter would be pawned off on the unsupported modem and therefore be not their problem.



Right. The onion routers remember where your machine is and route packets accordingly (I'm being overly simplistic, I know). Is this not the same as network address translation? I mean, my router remembers where my machine is at and routes the packets that belong to it accordingly... as far as AT&T knows (in my case), all data is being requested by and sent to one single IP, the router. It's the router's job to figure out which packets belong to my machine and which belong to the Playstation, or to the television.



I certainly will do so.

One thing. If you type your router's IP address into the address bar, it will show you all the devices connected to the router and how.

[Dngrsone]

Originally Posted by Cinisajoy

One thing. If you type your router's IP address into the address bar, it will show you all the devices connected to the router and how.

Can AT&T see that from their side of the router?[/quote]
I don't believe so, and I see no reason why they would care. They have better things to do than dig down into stuff like that.

They provide an IP address to you, which is bound to your router. What's behind your router is blocked off by the router.
______
Dennis

[Cinisajoy]

Originally Posted by Dngrsone

Can AT&T see that from their side of the router?

Yes, they can.
I know the AT&T tech guy was shocked when he saw 10 or so devices. Now they can't always tell what they are.

Amusing note when we switched to Uverse and the second tech came out, he refused to connect my devices for me. He figured with the laptops and the tablets I knew how.
Though he did reconnect all the ethernet stuff.
Oh and telling him but the tech did my parents stuff just led to giggles.

Second tech because the first one refused to run a new line saying that wasn't covered. And due to running the phone cord across the ceiling and down to the modem/router, it was having fits at times.
Yes, it was covered to have the new line ran.
Now it works great unless someone hits a telephone or electric pole.

[DMcCunney]

Originally Posted by Cinisajoy

One thing. If you type your router's IP address into the address bar, it will show you all the devices connected to the router and how.

Doesn't work that way here. If I go to my router's IP address, I get an HTML page generated by the router that is the management interface. From there, I can dig down to see what is connected and what IP address the router provided it, but that's an extra set of steps. I don't see it just by typing the router's address into my browser, and I've never seen a router that does work that way. What make/model do you have?
______
Dennis

[Cinisajoy]

Originally Posted by DMcCunney

Doesn't work that way here. If I go the my router's IP address, I get an HTML page generated by the router that is the management interface. From there, I can dig down to see what is connected and what IP address the routed provided it, but that's an extra set of steps. I don't see it just by typing the router's address into my browser, and I've never seen a routher that does work that way. What make/model do you have?
______
Dennis

Oh duh, me idiot.
You have to make sure you are in the device tab and click device info.

Motorola NVG510.

[Dngrsone]

Hrm... well, with the firewall appliance in place, I'm thinking they won't see anything on my side of the firewall. Kind of the whole purpose.

[DMcCunney]

Originally Posted by Dngrsone

Hrm... well, with the firewall appliance in place, I'm thinking they won't see anything on my side of the firewall. Kind of the whole purpose.

Yes.

My combo cable modem/Wifi router is secured with WPA2/PSK, and the firewall is enabled. I also have remote administration turned off, since I have no need to do so. If I need to make adjustments on my router, it happens from my desktop, which doesn't have Wifi and connects via a CAT5 cable. If I connect to do admin, I must supply the admin password.

For my ISP to be able to see what connects to my router, I'd have to turn off security, enable remote admin, and let them in. I have no reason to do so. The past couple of times I had issues with the modem I needed to talk to my ISP about, the first case was my original modem finally failing and needing replacement. The second was a needed firmware upgrade. But all the tech could see from his end was the modem itself and its status. He could not see what was behind it, and had no need to. He could talk me through what was needed.

Before I got the new combo modem/router, the cable modem connected to an external router. My router of choice was a Linksys WRT54G. That model used a Linux kernel, and because it did, the firmware was open source and could be modified. Various folks did, and my third party firmware of choice was a product called Tomato. Tomato had a vastly improved HTML interface to router functions with a finer degree of control, and would let me SSH into the router to get to a command line. My SO was bemused that I could run vi on the router to diddle config files.

The Linksys finally failed, and the later WRT54G models dropped the amount of installed memory and switched to a different OS kernel that was unhackable. To get a Linksys with Linux firmware, I'd have to get the WRT54GL model, at a higher price.

I've been thinking about getting an older Linux WRT54G off of eBay and running Tomato again. I'd put the TWC supplied Arris modem/router into bridge mode and let the WRT54G handle DHCP and routing. It's not a pressing issue because the Arris is configurable enough. It's mostly because I like to play.
______
Dennis

[DMcCunney]

Originally Posted by Dngrsone

Not really. I mean, as I said, one terminal. The computer was being used to operate an array of test equipment. There were a handful of subroutines that had to be running in the background (we're talking tens of individual components ranging from power supplies, measurement devices and switching to link it all together, using a half-dozen communications busses), but in the end, thre was only one user on the computer at any particular time.

There may have been only one user, but there was more than one job. Those background routines were started by something. On a *nix system, the something will be a system ID like root or bin which will be considered the user. I don't know how the Harris box did things.

I'm thinking that Harris found an excuse to offload some excess inventory and still fulfill their contract.

If this was a government contract, no surprise. The contracting officials will release an RFP, vendors will bid on the contract, and the winner will do what the RFP calls for, but it tends not to specify how it is done, only the results desired. If the vendor can unload gear while staying within budget, so much the better. (And in some cases, vendors will submit an unrealistically low bid to win the contract, assuming they can go back to the well for more funding, because it's easier for the contracting agency to give them more money than to fire them and have a different contractor restart from scratch.)

Same here: there was a card reader interface, and theoretically there were card readers available (actually in one facility there was one on-hand, just not in my shop), and our software programs were written and structured for storage on punched cards, but really...

The card reader in my shop was for the odd emergency where a job might need to be loaded from a card deck, but I never heard of it getting used. Even though storage and structure was for punch cards, the cards had gone virtual.

One of my instructors for the bench told us that our job (the one we were studying for) was a Cold war relic: should there be an nuclear blast (and accompanying EMP) that wiped out the data on our removable discs (those aforementioned 18" platters), then they could send out a card reader and stacks of cards and we could recompile all our programs and get running again. It sounded logical... in a Hollywood kind of way. Realistically, though, if the EMP were enough to penetrate several feet of steel to get to our platters and wipe them, then all the electronics in the entire test station would be crap, so they'd be shipping an entire bench or two, plus the software discs take up a lot less room and weigh less than an equivalent amount of punched-card data.

There are similar notions embedded all over. The Internet and TCP/IP had roots in a DARPA project to create a distributed fail-safe network that could route around failed nodes. The assumption in the Cold War times in which the initial design was done was that a node might fail because it had been the target for a nuclear strike.

I've seen doomday comments on the effects of Google going down, and all I can say is "If a disaster of a magnitude to put all of Google down occurs, being unable to reach Google will be the least of your problems..."

More likely is that Harris sold the card reader module with the computer or sold the government on the necessity to get rid of some of that excess inventory.

Might not have been excess inventory. But if your business is in part selling hardware, and you can sell some to the government, you do so. For all I know, having an actual card reader "just in case" might have been a contract requirement.

Hahaha. Yeah, I learned a lot more than most because when one of our guys left the shop he slipped a little something into the benches to mess with us: whenever someone loaded up a specific program (for the easiest thing we had to work on), the bench would essentially tell the tech to quick cherry-picking and go work on something important.

It took a little research, but I finally figured out and found the script he wrote. It resided on the system disc and since the system disc would be searched before the removable drive, his script would execute first. The script was basically print message, go to removable drive and execute.

I saw tricks like that played on the IBM system.

It's the same as substituting a custom script for an actual bash command by placing your script in a higher-priority directory.

If the system is properly set up, you can't do that. Among other things, scripts have absolute path names specifying the name and location of the interpreter that will run the script as the first line in shbang format, like "#!/usr/bin/perl", and if the script is running an external command, you can specify it by name and location. And if you're smart, you explicitly set $PATH in your script and include only the default system directories, so someone can't substitute something malicious with the same name that lives elsewhere. Unlike MSDOS and Windows, *nix only looks in the defined PATH when you specify a command to run. They do not look in whatever the current directory happens to be first the way DOS and Windows do. To get them to look in the current directory at all, you must explicitly add "." to your PATH, but you add it as the last entry so the system dirs all get searched first.

Exactly. The hack is basically exploiting a vulnerability on the modem and running commands as root, then using its nsh console to set it to true bridge mode, which is not possible through the public shell, by redirecting the PPP stream through one of the ethernet ports. reference

<sigh> A decent modem/router shouldn't require you to exploit a vilnerability to do that. It should be an option in the administrative interface.

Been there, done that: Palm T|X for close to eight years before I was forced to use a 'smart' phone.

This wasn't a TX. It was a Tapwave Zodiac 2, and you could get a Wifi SD card to plug in to get it on line. That card only supported WEP.

The TX had a firmware update from Palm that added WPA authentication, and mine has it applied, but I've never been able to get it to connect to my router.

Going online with the Zodiac was mostly a "because I could" exercise, used seldom. (I did test that I could SSH from the Zodiac at home to a *nix server in the office, but I dreaded having to actually do it for anything serious.)

Right. The onion routers remember where your machine is and route packets accordingly (I'm being overly simplistic, I know). Is this not the same as network address translation? I mean, my router remembers where my machine is at and routes the packets that belong to it accordingly... as far as AT&T knows (in my case), all data is being requested by and sent to one single IP, the router. It's the router's job to figure out which packets belong to my machine and which belong to the Playstation, or to the television.

I suppose you can think of it that way. When you send a TCPIP packet, the packet header has a source IP and destination IP address. If you use Tor, you don't want your real IP address so Tor diddles the packet headers internally. The idea is that the system you connect to doesn't know the true origin of the traffic and can't find out.

On a home router, NAT is used to have multiple devices behind your router connect through one public IP. The router maintains connection and state information, so it knows what packets coming back are intended for what device in the internal network and can pass them along.

I certainly will do so.

I look forward to it.

[Dngrsone]

Originally Posted by DMcCunney

<sigh> A decent modem/router shouldn't require you to exploit a vilnerability to do that. It should be an option in the administrative interface.

Well... when's the last time a big name ISP sold a decent modem/router to the consumer market?

And now with the new FCC rules, I doubt that any consumer-grade modem/router will allow custom firmware installation.