- NANOG - lists.nanog.org

Re: Filtering (was Re: RFC 1918)
by Paul Vixie 19 Jul '00

19 Jul '00

Valdis.Kletnieks(a)vt.edu writes: > OK.. I'll bite - what's in YOUR favorite list of things to filter? ;) networks owned by spam-friendly people or companies.

1 0

Re: RFC 1918
by Richard A. Steenbergen 19 Jul '00

19 Jul '00

On Mon, 17 Jul 2000, Eric A. Hall wrote: > When ISPs choose to mark their packets with Internet-illegal addresses, > they are contributing to these problems. Sorry, but you're not supposed > to be using these addresses anyway. This is utterly stupid. You can use these addresses any way you see fit, you can source packets from them if you'd like, and they are as valid as any other address to use and be "on the internet". What you CAN'T do however, is expect that these packets can ever be replied to, or exchange or accept any kind of information on how to route this IP space outside of your network. Sure its probably not the best idea in the world to send out packets you can't expect a reply to, but its not prohibited for a reason, and its certainly not the end of the world you make it out to be. If you really want to filter RFC1918 sourced packets at your borders for whatever reason its your choice. Trust me I've probably seen a lot more DoS then most people in one way or another, and filtering 1918 space is in absolutily NO way any kind of magic bullet or even worth the processor time (if you're gonna spend the time filtering there are much better things out there). -- Richard A Steenbergen <ras(a)e-gerbil.net> http://www.e-gerbil.net/humble PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)

3 2

Re: RFC 1918
by Bohdan Tashchuk 18 Jul '00

18 Jul '00

> Though technically you're right, this kind of attitude is exactly the > problem. Everyone should filter all RFC1918 usage on public links, > regardless of whether they themselves use is, or their customers use it, > or not. To not do such filtering is to be a bad neighbour. I'm just a home DSL user, and usually a fly on the wall for this mailing list. But I've found it beneficial to make a practical exception to your blanket condemnation of RFC1918 source addresses. I don't think there is much harm to it. The relevant snippet of my rules on my ingress filter is: 1) ... block bad things such as unused or spoofed addrs ... 2) allow icmp from any to any icmptypes 0,3,4,11,12 3) deny ip from 10.0.0.0/8 to any 4) deny ip from 172.16.0.0/12 to any 5) deny ip from 192.168.0.0/16 to any 6) allow tcp from any to any 1024-65535 established 7) ... some other rules ... 8) deny everything else by default Line #2 allows relatively benign incoming ICMP, such as "fragmentation needed", but hopefully blocks the more problematic stuff. I added this exception for a very practical reason. Without it there were many routers, generating ICMP messages using RFC1918 source addresses, whose error messages were important, but that I dropped at the firewall. Interestingly, these messages passed thru MANY intermediate routers that didn't block packets with RFC1918 source addresses. If you take it upon yourself to "filter all RFC1918 usage" from the outside world, you (and your customers) will suffer for it. Because it seems to be established practice out there. Of course I never send packets to the Internet with an RFC1918 address in them.

7 9

ARIN Meetings and Upcoming Elections
by Richard Jimmerson 18 Jul '00

18 Jul '00

ARIN will hold it's biannual public policy and members meetings on October 2 through 4 in Herndon, Virginia. Registration and other information about these meetings can be found under the announcements section of ARIN's front page (http://www.arin.net) Information about upcoming ARIN elections can also be found under the announcements section of ARIN's front page. There is currently an open call for ASO AC election nominations and member volunteers are being sought for the BOT elections nominating committee. Updated information regarding ARIN elections will be announced and posted to ARIN's website periodically between now and October. Regards, American Registry for Internet Numbers (ARIN)

1 0

What's a termination
by Sean Donelan 18 Jul '00

18 Jul '00

On Mon, 17 July 2000, "Patrick W. Gilmore" wrote: > The telco world is completely different. *EVERY* call has an "origination" > and a "termination", and *EVERY* termination is *PAID*. Every "telco" termination is paid. However the telco's like to define some terminations as other than a telco, so the telco doesn't pay to terminate those calls. For example, in the USA calls to and from cellular carriers are charged access fees by the wireline carrier. Telco's first wanted to charge access fees for calls to ISPs, now they want a "Free-Ride"(tm) to terminate calls without reciprocal compensation to places they don't want to pay. I think a big difference is in the telco world, telco's collect more money for more call minutes they complete. If the call isn't completed, they don't get money. So telco's have an incentive to complete (and bill) as many call minutes as possible. Much of the US Internet is still flat-rate, so ISPs collect the same amount of money for most levels of service good or bad. ISPs have an incentive to deliver the least quality of service they can before losing customers. Even if an ISP has a multi-day outage, they bill for the service unless customers get so irate they may receive a token refund. Even ISPs with SLAs tend to make the refund procedure so cumbersome, they know its like giving out coupons instead of lowering the price. Only a small number of customers will ever get through the process and collect.

1 0

Telco Network Operations Forum
by Sean Donelan 18 Jul '00

18 Jul '00

On Tue, 18 July 2000, Robert Cannon wrote: > I am curious how often you think that ATT telephone long distance would hand > traffic off to MCI telephone long distance. My impression of telephone long > distance is that it is largely an end to end service (one of the great > differences from the Internet). That as far as long distance goes, there is > not a great deal traffic hand off (one exception is times of network trouble > where carriers have agreements to hand off traffic to maintain network > reliability). The urban legend has been MCI was AT&T's third largest customer (after General Motors and US Government). However as far as I know, there is no publicly available source of which telco uses what. Telco's swap a lot of things to reach different places. There are 1800 local telephone companies, but IXCs like Sprint or Worldcom only have a few hundred POPs around the country. So frequently an ILEC (usually the former RBOC) provides tandem service between an IXC and another LEC. Or an IXC may backhaul the call across another IXC's circuits. Look back to many of the "cream-skimming" complaints. Add in complications like Illuminet's SS7 service between multiple carriers, and a simple problem can involve a lot of different carriers. As Patrick mentioned, international calling is even more likely to involve "hidden" carriers. An inter-exchange carrier may not have its own international agreements with every carrier in every country. So if you are in PacBell land with Qwest as your long distance provider and call some place like Cuba it may pass through several other carriers or countries on its way to Cuba. And for odd requests, such as a person-to-person call to a sub-sahara country, the IXC will punt the call and tell you to call AT&T for help. I've heard of one international circuit between New York and Stockholm involving 13 different carriers. And no, the telephone carriers don't speak to each other any better. ATIS has an operations group which maintains an extremely thick manual how telco network management and control centers are supposed to work together. If you read the minutes from their meetings, it sounds remarkably like polite NANOG. The problems are the same, the words just have more syllables. Read Cliff Stoll's book The Cuckoo's Egg for a fun description of how many telephone companies may be involved in a call as he tried to trace his intruder. The National Communication System compiled a database covering the physical connectivity of the three major IXCs and dominant ILECs under an non-disclosure agreement for "national security/emergency preparedness" purposes. But even that is incomplete. And woefully inadequate for trying to predict what would happen to the Internet because most Internet circuits seem to run on the new carriers with cheap bandwidth instead of the three major IXCs. For example, UUNET uses some non-MCI/Worldcom circuits, as you can often see when Worldcom has a fiber cut isolating the Worldcom POP, but UUNET is unaffected.

1 0

Re: Telco NOC vs. Internet NOCs (WAS: Wanted: Clueful Individual @ TeleGlobe.net
by Robert Cannon 18 Jul '00

18 Jul '00

I am curious how often you think that ATT telephone long distance would hand traffic off to MCI telephone long distance. My impression of telephone long distance is that it is largely an end to end service (one of the great differences from the Internet). That as far as long distance goes, there is not a great deal traffic hand off (one exception is times of network trouble where carriers have agreements to hand off traffic to maintain network reliability). Also sprach Patrick W. Gilmore >The telco world is completely different. *EVERY* call has an >"origination" and a "termination", and *EVERY* termination is *PAID*. Example 1: User A goes off hook to call user B who is in another state. User A's local carrier is Bell South. Long distance is Qwest. User's B local carrier is PAcBell. User A pays a bill to Bell South and Qwest. Qwest pays an origination and termination access fee to Bell South and PacBell. User B pays PacBell. So there are flat rate charges on local networks, and metered origination and termination fees on the long distance network. Example 2: User A goes online to "webphone" and using A's webbrowswer calls User B. User's A Net access is not over the Public Telephone Network. "Webphone" is a free service in affiliation with GTE. User B subscribes to PacBell. User A pays a flat rate bill to its Net access provider; user A pays not long distance fee. GTE could act as a CLEC or a long distance company. If it acts as a long distance company IXC then GTE pays a termination access charge but NO origination access charge! User B pays a bill to pacbell. Example 3: User A calls User B 100% over the net and there are no telco charges. Example 4: User A calls User B. Both are on the Sprint Wireless network. Both pay flat rate charges to Sprint for 300 minutes of time. Since the entire call is on network, there are no access charges and reciprocal compensation. If at one end there are two local networks passing traffic, then the local networks must pay a metered reciprocal compensation charge. There are also universal service fees involved. Point: You cannot say "every" call has certain fees. The telephone network gets more diverse by the moment. If there is a trouble with the service in the pacbell network, what exactly do you want bell south to do about it? I guess I an not sure I understand. Finally, dont hestitate to drop essentially an email to the FCC letting them know trouble that you are having with your service. http://www.fcc.gov/eb/tcd/complaints.html Thanks Robert Cannon Internet Telecom Project www.cybertelecom.org ------Original Message------ From: Jeff Mcadams <jeffm(a)iglou.com> To: nanog(a)merit.edu Sent: July 18, 2000 12:27:11 PM GMT Subject: Re: Telco NOC vs. Internet NOCs (WAS: Wanted: Clueful Individual @ TeleGlobe.net Also sprach Patrick W. Gilmore >The telco world is completely different. *EVERY* call has an >"origination" and a "termination", and *EVERY* termination is *PAID*. ______________________________________________ FREE Personalized Email at Mail.com Sign up at http://www.mail.com/?sr=signup

2 1

ARIN to Allocate from 65.0.0.0/8 and 66.0.0.0/8
by Richard Jimmerson 18 Jul '00

18 Jul '00

In the near future, ARIN will begin making allocations from 65.0.0.0/8 and 66.0.0.0/8. This will include allocations of /20 and shorter prefixes, according to ARIN's minimum allocation policy. For informational purposes, a list of ARIN's currently administered blocks can be found at http://www.arin.net/regserv/IPStats.html Regards, American Registry for Internet Numbers (ARIN)

1 0

Telco NOC vs. Internet NOCs (WAS: Wanted: Clueful Individual @ TeleGlobe.net)
by Patrick W. Gilmore 18 Jul '00

18 Jul '00

At 10:14 AM 7/17/00 -0700, Karyn Ulriksen wrote: > >Out of curiousity, how does the telco world handle this? When some one >signs up for MCI longdistance, I assume that on some calls they will cross >off the MCI network (and potentially though several others) that will be >piled mile high in static. The customers will complain about their >connection to their respective local and long distance companies. But I >believe it isn't an option to contact the networks that they may transit >mid-stream. However, some QC must be in place to achieve some level of >consistancy to get that expected "pin drop" quality we've apparently come to >take for granted... The telco world is completely different. *EVERY* call has an "origination" and a "termination", and *EVERY* termination is *PAID*. So when it goes from you (user) to AT&T to MCI to Telstra to friend (user), you pay AT&T, AT&T pays MCI, and MCI pays Telstra. (I made that sequence up. I doubt a call would be routed like that, but you get the point.) So, if there is a problem inside MCI, AT&T will hound them about it. Even though no "user" is paying MCI, AT&T *is* paying them, and has every right to get upset if quality is not up to par. Plus, the user has no idea of routing in the phone system. There is no traceroute, there are no in-addrs, etc., etc. As a user, you call AT&T, and that is all you know. But that is okay, because AT&T takes full responsibility for end-to-end connectivity. If the call does not get through, they may say "MCI is having a problem", but they will work on it for you. Some may say this is why the telcos work better than the Internet. Personally, I think it is kinda silly. The way the system is laid out, it costs a lot more to bill a call than connect a call. It takes a lot more to track the minutes of a call and all the various tariffs, interconnections, time-of-day billing, billing cycles, interest / late fees, mail out a bill every month, etc., etc., etc., than it takes for a few switches to route a call over fiber which costs nearly nothing anyway. Remember, typical pricing for trans-continental US fiber, when buying dark or very large IRUs, is a few places to the right of the decimal point per month per DS0-mile - or *at most* $3/month for a cross country DS0, and for most big carriers it is orders of magnitude less. And the telcos tend to get a wee bit more than $3/month revenue off that DS0. ;) Of course, there is cap-ex for the switches & fiber lighting equipment - which is a huge expense. Plus people to operate it and stuff like that. But the point is, it costs more to bill it than to actually *do* it. I do not think we want to see the Internet go that way. Do you? I am not saying this is a valid way to run the Internet, just that it is not how I would like to see the Internet run. Perhaps my view / opinions are silly or impractical, but that is how I feel. What do all of you think? >Karyn TTFN, patrick

2 1

RE: Telco NOC vs. Internet NOCs (WAS: Wanted: Clueful Individual @ TeleGlobe.net)
by rdobbins＠netmore.net 18 Jul '00

18 Jul '00

You're certainly correct in your explanation of tolls, tariffs, etc.; however, do bear in mind that the telcos are heavily regulated, that they have quality-of-service and 'uptime', if you will, standards they must meet from the various agencies who impinge upon them (both Federal and local), and that they also must get both price increases -and- decreases approved at the various affected levels, as well. All these things are part of what contribute to the Soviet-style mentality found within most telcos. They're also part of the reason that - with very few exceptions - telephone service from the user perspective works far more reliably than Internet connectivity. I personally would never want the Net to go the way of the telcos, after having worked for one. Conversely, there -are- some lessons about quality-of-service which the ISP world can learn from them. My greatest fear in this regard is that as Internet connectivity becomes something that everyone expects and needs, just like their telephone service (or even becomes their telephone service), that the regulators will step in, unless we all can get our acts together enough to ensure that they've nothing to worry about. I'm not naive enough to think that they won't try to regulate our operations, anyway, but if we do our best to get our own house in order, they'll have less of a change of actually succeeding. ----------------------------------------------------------- Roland Dobbins <rdobbins(a)netmore.net> // 818.535.5024 voice -----Original Message----- From: Patrick W. Gilmore [mailto:patrick@ianai.net] Sent: Monday, July 17, 2000 9:07 PM To: nanog(a)merit.edu Subject: Telco NOC vs. Internet NOCs (WAS: Wanted: Clueful Individual @ TeleGlobe.net) At 10:14 AM 7/17/00 -0700, Karyn Ulriksen wrote: > >Out of curiousity, how does the telco world handle this? When some one >signs up for MCI longdistance, I assume that on some calls they will cross >off the MCI network (and potentially though several others) that will be >piled mile high in static. The customers will complain about their >connection to their respective local and long distance companies. But I >believe it isn't an option to contact the networks that they may transit >mid-stream. However, some QC must be in place to achieve some level of >consistancy to get that expected "pin drop" quality we've apparently come to >take for granted... The telco world is completely different. *EVERY* call has an "origination" and a "termination", and *EVERY* termination is *PAID*. So when it goes from you (user) to AT&T to MCI to Telstra to friend (user), you pay AT&T, AT&T pays MCI, and MCI pays Telstra. (I made that sequence up. I doubt a call would be routed like that, but you get the point.) So, if there is a problem inside MCI, AT&T will hound them about it. Even though no "user" is paying MCI, AT&T *is* paying them, and has every right to get upset if quality is not up to par. Plus, the user has no idea of routing in the phone system. There is no traceroute, there are no in-addrs, etc., etc. As a user, you call AT&T, and that is all you know. But that is okay, because AT&T takes full responsibility for end-to-end connectivity. If the call does not get through, they may say "MCI is having a problem", but they will work on it for you. Some may say this is why the telcos work better than the Internet. Personally, I think it is kinda silly. The way the system is laid out, it costs a lot more to bill a call than connect a call. It takes a lot more to track the minutes of a call and all the various tariffs, interconnections, time-of-day billing, billing cycles, interest / late fees, mail out a bill every month, etc., etc., etc., than it takes for a few switches to route a call over fiber which costs nearly nothing anyway. Remember, typical pricing for trans-continental US fiber, when buying dark or very large IRUs, is a few places to the right of the decimal point per month per DS0-mile - or *at most* $3/month for a cross country DS0, and for most big carriers it is orders of magnitude less. And the telcos tend to get a wee bit more than $3/month revenue off that DS0. ;) Of course, there is cap-ex for the switches & fiber lighting equipment - which is a huge expense. Plus people to operate it and stuff like that. But the point is, it costs more to bill it than to actually *do* it. I do not think we want to see the Internet go that way. Do you? I am not saying this is a valid way to run the Internet, just that it is not how I would like to see the Internet run. Perhaps my view / opinions are silly or impractical, but that is how I feel. What do all of you think? >Karyn TTFN, patrick

1 0