I replaced Netscape with Firefox's first released version, even gave money to the Mozilla foundation back then. Tab Groups became my favorite feature, then after they trashed it switched to the add on. Now they've killed that ability and the developer gave up.
Lets see, regular hangs and crashes but with tab groups, or a regularly hanging and crashing Chrome clone without them. Hmm, tough decision. Guess it's bye bye to the fox and move to our Googlian Overlords tool of global domination the day my tabs all become flat again. Just have to get used to all the social engineering warnings, really don't care if the NSA sees what I read on/. or if a major forum uses a cert that isn't exactly perfect per our overlords rules
35 +/- years. Gosh, must be getting old.
on
30 Years a Sysadmin
·
· Score: 1
Was a mainframe systems programmer on IBM and Univac machines starting about 1972 or so.
First Unix machine I worked on was a Callan workstation somewhere around 1981, followed by 68K Suns, Motorola VME buss systems (which we built from spare parts and installed Sys V on), Sparc Suns, then the usual Linux suspects somewhere around 99 or 2000. Sysadming wasn't my primary job, but somehow I was in charge of keeping them going.
I did see a PDP-11 Unix machine in the mainframe room at Western Electric Greensboro, used the console's desk as a footrest while running tests on IBM mainframe front end systems in the middle of the night somewhere in the mid 70s. Asked, but they wouldn't let me fiddle with it.
From a quick RTFA the initial user has a DLink router.
FWIW, I bought a DLink wireless router a year or so back for my home network, don't recall the model, that would not do WPA2/TKIP with Windows (yeah, I know) Vista or XP, or my PSP. I'm an experienced network engineer, not a novice. It took a couple days fooling with it, several support emails, and then several hours on the phone with DLink before they finally said WPA was broke and to use WEP. IIRC Windoz was logging authentication errors.
The DLink got returned and replaced with a Netgear WGR614 that worked the first time, and still works today.
I ran the network for a regional ISP in low population density Michigan's Upper Penisula., taking it from 0 to 18000 customers, before retiring and changing careers. Being an ISP means that you rent or build the road, not the car. You have to provide for the peak bandwidth, or as close as you can get.
Back in the pre-broadband days we had a bit under 2000 dial lines, and the equivalent of a P2P bandwidth hog was somebody who stayed dialed in 24 hours a day - we called them port hogs and they were 3 to 5 percent of our total customers (that 3 to 5 seems to be the problem customer percentage in almost any business). Since they paid $20/month, tied up a port costing us $100/month (line, routers, short lifetime modems, backhaul, overhead, etc.), and our average margin was about a buck/customer/month. We sent them a nice letter asking if they'd please hang up when they were away from the computer/home - that worked about half the time. Then we automatically disconnected them every two hours, which gave about half the rest the idea. The holdouts got a nice letter saying that they were going to receive out best service because of their extraordinary needs, and to confirm the upgrade they just had to keep their usage up. We'd give them their own private phone number, and a bill for $150/month. Only one customer ever got this upgrade. Most of the rest went to our competitor - who was close to bankrupt because he couldn't keep up his phone payments. Buck a month isn't much room for error. Our threshold, btw, was 8 hours a day, 16 on weekends. If you work and sleep you can't use that much anyway and our basic dial TOS was unlimited interactive, defined as sitting at the computer using it. Average customer IIRC was 50 hours/month when the ultra high timers weren't counted, and at 10:1 it was a rare night that got busy signals and usually for just a few minutes. Our occasional high busy times were almost all busted boxes or phone problems.
Word got around - we lost some business, but attracted a lot of customers because we had the fastest network and fewest busies in the area. Monitoring bandwidth and dial use counts and trying to beat the projections during the several month leadtime didn't hurt.
Broadband's a bit different - no worrying about busies - but you need to buy a lot more bandwidth. Normal subscriber rates don't even come close to the cost of a full pipe 100% of the time from the POP so it has to be shared. The hard core P2Per is trying to get hundreds or thousands of dollars worth of service for his $30 (or whatever) a month. Since they lock the doors if you don't at least break even you can either cap and bill, limit bandwidth, or have them go to the competition and hurt their margins (which are still pretty small).
So yeah, I understand peering too. ISPs do get charged for every bit they have to get from that peering point to the customer and somebody always has to pay for the pipes to make it all work - no free lunches here and P2P's problem is the increased bandwidth everywhere. Peering payments are to correct the imbalance between what you've paid to haul your users traffic to a peering point (it's in a fixed location, after all) for his customers vs what the other guy's paid to haul his customers to yours.
Was it Verizon who was trying to tweak bittorrent so the closest peer was the most used? Fair something? A great idea, particularly if both ends are on the same backroom ethernet switch.
Not everyone uses akamai. And like I said someone pays the bills one way or another even if you directly don't pay. Why's that a hard concept for this forum to grasp?
Makes you wonder how many of them work and support themselves...
It's hard to grasp because it's wrong.
Peering does not necessarily involve payment between the parties. A and B agree to exchange traffic. What goes from A to B isn't subject to a monthly fee, just like what goes from A to A doesn't. It might require bigger pipes at some point but it's not metered.
You live in a peering point with all your content providers? "Free" peering still needs a way to get to and from the NAP. The internal pipes cost big bucks as soon as they use the phone company, metered or not. Price a thousand km OC48. If the ISP isn't a charity with very deep pockets or tax-money-spender (which means they're spending *your* money very inefficiently) either the customer pays or you're in the bankruptcy court.
BTW, peering agreements just mean we'll call the traffic equal, your customers using as much of mine as mine use of yours, rather then billing each other the same amount.
Put another way, you go to the store to bring home a fine bottle of wine. You give the guy at the door an equivalent bottle. That's "free" peering. You try to give him a bottle of Ripple he's not going to be happy. That's trying to peer with unequal traffic.
7 hospitals!
We cover an area in Upper Michigan around 60 miles north to south, 30 E-W (it's a peninsula) with two hospitals that have level 4 (working on 3) ERs, with a 100 mile transport to the nearest level 3 (working on 2) trauma center. The nearest helicopter is Duluth, 45 minutes away on a good day.
I do a couple 100 mile transfers, mostly medical, every week. Average response is probably around 20 minutes, up to 45-60 in the way out areas in the winter (we average 200+ inches of snow a year). Our service does 1100 local and 500 transfers a year. Medic and EMT crews, one at night, two during the day.
We had a wreck a couple weeks ago while I was on night shift, rollover into a power pole at 90-100 mph, with 5 kids. Three ejected, one DOS, two critical, and the two with seat belts were minor. Two made the 100 mile trip, one dying in surgery. My kids knew one of the boys that died.
My last shift was a nap, lunch, another nap, and a minor medical call. Pretty typical - lots different than where you are...
I'm also a paramedic, with a rural ambulance service. Started as a Medical First Responder, when the local fire department chief got his wife to get my wife to get me to take the class. I was running a local ISP from a telephone exchange building, and was one of the few who actually worked in our community rather than the "big" (5000 people) city 20 miles away.
Pacbot is made by the same people as Roomba. What happens when it does the famous Roomba Circle Dance?
For the single non-Roomba owner out there (sorry; it's still on backorder), Roombas are prone to sensor problems - usually caused by dirt - that makes them start going in circles.
Imagine a spinning Roomba wildly firing a shotgun...
You're right, I typed too fast. There was 64 sectors of 28 36 bit words per track, giving 1792 words/track. 28 is, of course, a perfect number.
Ours were sitting on steel I-beams directly on the concrete floor - otherwise they'd crack the floor.
The 1107 was not the Univac I, nor was the architecture similar.
The story goes that Univac numbered their machines from 1 to 12. They didn't want to call something the Univac 13, so they named it the 1101, or 13 in binary. The 1101 evolved into the 1107, 1108, etc, and still lives on as the 2200 series.
The Univac FASTRAND II drum, supposedly made of machined sewer pipe, was used on the early 1100 machines. It had 192 positions of 64 tracks of 64 sectors of 36 bit words each, and was the primary storage device on the 1107s I worked on in the early 70s.
It's not painless for the ISP; their stated cost per customer service call is $9.00. From my years in the ISP business I'd say it's not unreasonable cost estimate, either.
Plan B: make it automatic. Have the user log in somewhere with their Comcast email address/password (that'd get rid of the clueless users, who don't even remember that they *have* a password, let alone what it is) and say they need port 25 opened. Comcast should still monitor volume on outbound port 25.
And for the tinfoil hat crowd, what's so bad about relaying email through an ISP mail server? If it's sensitive encrypt it, since Carnivore will read it anyway...
Slashdot Mirror
I replaced Netscape with Firefox's first released version, even gave money to the Mozilla foundation back then. Tab Groups became my favorite feature, then after they trashed it switched to the add on. Now they've killed that ability and the developer gave up. Lets see, regular hangs and crashes but with tab groups, or a regularly hanging and crashing Chrome clone without them. Hmm, tough decision. Guess it's bye bye to the fox and move to our Googlian Overlords tool of global domination the day my tabs all become flat again. Just have to get used to all the social engineering warnings, really don't care if the NSA sees what I read on /. or if a major forum uses a cert that isn't exactly perfect per our overlords rules
Was a mainframe systems programmer on IBM and Univac machines starting about 1972 or so. First Unix machine I worked on was a Callan workstation somewhere around 1981, followed by 68K Suns, Motorola VME buss systems (which we built from spare parts and installed Sys V on), Sparc Suns, then the usual Linux suspects somewhere around 99 or 2000. Sysadming wasn't my primary job, but somehow I was in charge of keeping them going. I did see a PDP-11 Unix machine in the mainframe room at Western Electric Greensboro, used the console's desk as a footrest while running tests on IBM mainframe front end systems in the middle of the night somewhere in the mid 70s. Asked, but they wouldn't let me fiddle with it.
The worse threat to ham radio is not cussing and https, but to simple lack of amateurs
Actually, the number of US amateur radio licenses reached an all-time high last year.
http://www.arrl.org/news/2012-marks-all-time-high-for-amateur-radio-licenses
kc8mmu
If it doesn't take two boxes of cards it's not a real program.
From a quick RTFA the initial user has a DLink router.
FWIW, I bought a DLink wireless router a year or so back for my home network, don't recall the model, that would not do WPA2/TKIP with Windows (yeah, I know) Vista or XP, or my PSP. I'm an experienced network engineer, not a novice. It took a couple days fooling with it, several support emails, and then several hours on the phone with DLink before they finally said WPA was broke and to use WEP. IIRC Windoz was logging authentication errors.
The DLink got returned and replaced with a Netgear WGR614 that worked the first time, and still works today.
I ran the network for a regional ISP in low population density Michigan's Upper Penisula., taking it from 0 to 18000 customers, before retiring and changing careers. Being an ISP means that you rent or build the road, not the car. You have to provide for the peak bandwidth, or as close as you can get.
Back in the pre-broadband days we had a bit under 2000 dial lines, and the equivalent of a P2P bandwidth hog was somebody who stayed dialed in 24 hours a day - we called them port hogs and they were 3 to 5 percent of our total customers (that 3 to 5 seems to be the problem customer percentage in almost any business). Since they paid $20/month, tied up a port costing us $100/month (line, routers, short lifetime modems, backhaul, overhead, etc.), and our average margin was about a buck/customer/month. We sent them a nice letter asking if they'd please hang up when they were away from the computer/home - that worked about half the time. Then we automatically disconnected them every two hours, which gave about half the rest the idea. The holdouts got a nice letter saying that they were going to receive out best service because of their extraordinary needs, and to confirm the upgrade they just had to keep their usage up. We'd give them their own private phone number, and a bill for $150/month. Only one customer ever got this upgrade. Most of the rest went to our competitor - who was close to bankrupt because he couldn't keep up his phone payments. Buck a month isn't much room for error. Our threshold, btw, was 8 hours a day, 16 on weekends. If you work and sleep you can't use that much anyway and our basic dial TOS was unlimited interactive, defined as sitting at the computer using it. Average customer IIRC was 50 hours/month when the ultra high timers weren't counted, and at 10:1 it was a rare night that got busy signals and usually for just a few minutes. Our occasional high busy times were almost all busted boxes or phone problems.
Word got around - we lost some business, but attracted a lot of customers because we had the fastest network and fewest busies in the area. Monitoring bandwidth and dial use counts and trying to beat the projections during the several month leadtime didn't hurt.
Broadband's a bit different - no worrying about busies - but you need to buy a lot more bandwidth. Normal subscriber rates don't even come close to the cost of a full pipe 100% of the time from the POP so it has to be shared. The hard core P2Per is trying to get hundreds or thousands of dollars worth of service for his $30 (or whatever) a month. Since they lock the doors if you don't at least break even you can either cap and bill, limit bandwidth, or have them go to the competition and hurt their margins (which are still pretty small).
So yeah, I understand peering too. ISPs do get charged for every bit they have to get from that peering point to the customer and somebody always has to pay for the pipes to make it all work - no free lunches here and P2P's problem is the increased bandwidth everywhere. Peering payments are to correct the imbalance between what you've paid to haul your users traffic to a peering point (it's in a fixed location, after all) for his customers vs what the other guy's paid to haul his customers to yours.
Was it Verizon who was trying to tweak bittorrent so the closest peer was the most used? Fair something? A great idea, particularly if both ends are on the same backroom ethernet switch.
Not everyone uses akamai. And like I said someone pays the bills one way or another even if you directly don't pay. Why's that a hard concept for this forum to grasp?
Makes you wonder how many of them work and support themselves...
It's hard to grasp because it's wrong.
Peering does not necessarily involve payment between the parties. A and B agree to exchange traffic. What goes from A to B isn't subject to a monthly fee, just like what goes from A to A doesn't. It might require bigger pipes at some point but it's not metered.
You live in a peering point with all your content providers? "Free" peering still needs a way to get to and from the NAP. The internal pipes cost big bucks as soon as they use the phone company, metered or not. Price a thousand km OC48. If the ISP isn't a charity with very deep pockets or tax-money-spender (which means they're spending *your* money very inefficiently) either the customer pays or you're in the bankruptcy court.
BTW, peering agreements just mean we'll call the traffic equal, your customers using as much of mine as mine use of yours, rather then billing each other the same amount.
Put another way, you go to the store to bring home a fine bottle of wine. You give the guy at the door an equivalent bottle. That's "free" peering. You try to give him a bottle of Ripple he's not going to be happy. That's trying to peer with unequal traffic.
7 hospitals! We cover an area in Upper Michigan around 60 miles north to south, 30 E-W (it's a peninsula) with two hospitals that have level 4 (working on 3) ERs, with a 100 mile transport to the nearest level 3 (working on 2) trauma center. The nearest helicopter is Duluth, 45 minutes away on a good day. I do a couple 100 mile transfers, mostly medical, every week. Average response is probably around 20 minutes, up to 45-60 in the way out areas in the winter (we average 200+ inches of snow a year). Our service does 1100 local and 500 transfers a year. Medic and EMT crews, one at night, two during the day. We had a wreck a couple weeks ago while I was on night shift, rollover into a power pole at 90-100 mph, with 5 kids. Three ejected, one DOS, two critical, and the two with seat belts were minor. Two made the 100 mile trip, one dying in surgery. My kids knew one of the boys that died. My last shift was a nap, lunch, another nap, and a minor medical call. Pretty typical - lots different than where you are...
I'm also a paramedic, with a rural ambulance service. Started as a Medical First Responder, when the local fire department chief got his wife to get my wife to get me to take the class. I was running a local ISP from a telephone exchange building, and was one of the few who actually worked in our community rather than the "big" (5000 people) city 20 miles away.
Pacbot is made by the same people as Roomba. What happens when it does the famous Roomba Circle Dance? For the single non-Roomba owner out there (sorry; it's still on backorder), Roombas are prone to sensor problems - usually caused by dirt - that makes them start going in circles. Imagine a spinning Roomba wildly firing a shotgun...
Fluidics has been around for a long time.. http://en.wikipedia.org/wiki/Fluidic_logic
You're right, I typed too fast. There was 64 sectors of 28 36 bit words per track, giving 1792 words/track. 28 is, of course, a perfect number. Ours were sitting on steel I-beams directly on the concrete floor - otherwise they'd crack the floor.
The 1107 was not the Univac I, nor was the architecture similar. The story goes that Univac numbered their machines from 1 to 12. They didn't want to call something the Univac 13, so they named it the 1101, or 13 in binary. The 1101 evolved into the 1107, 1108, etc, and still lives on as the 2200 series. The Univac FASTRAND II drum, supposedly made of machined sewer pipe, was used on the early 1100 machines. It had 192 positions of 64 tracks of 64 sectors of 36 bit words each, and was the primary storage device on the 1107s I worked on in the early 70s.
Real geeks want an HP 16C replacement. My 16C's still going, and on only it's 3rd or so set of batteries since 1982. http://www.hpmuseum.org/hp16.htm
It's not painless for the ISP; their stated cost per customer service call is $9.00. From my years in the ISP business I'd say it's not unreasonable cost estimate, either. Plan B: make it automatic. Have the user log in somewhere with their Comcast email address/password (that'd get rid of the clueless users, who don't even remember that they *have* a password, let alone what it is) and say they need port 25 opened. Comcast should still monitor volume on outbound port 25. And for the tinfoil hat crowd, what's so bad about relaying email through an ISP mail server? If it's sensitive encrypt it, since Carnivore will read it anyway...
Sound a lot like freenet, if they ever solve the ongoing routing problems.
Can you find a better reference than a 12 year old academic paper?