|
Home | Reviews | Speed Test | Tools | News | Forums | Info | About | Join |
1.5 Interesting Info
No, the only list that was released to the public is more than two years old and hopelessly out of date. The list was originally meant to disclose to CLECs the overall plan but it has changed several times since it's release in 2000. by armyrebel4 edited by lev See this forum thread. "Redback" is the name of the company (Redback Networks), just like "Cisco" is used for Cisco Systems' products. These subscriber management systems (SMS) are their primary product line. Redback Networks' promo site is here. Their primary function is to consolidate and manage customer network traffic, authenticate users and aggregate thousands of individual connections into high speed pipes. by RadioDoc Telecommunications Act of 1996 •Sec 271 •Sec 272 Paragraphs (b), (c) and (e) •Sec 702 Amendment Title II inserted after section 221 (47 U.S.C.221) •Sec 222 Paragraphs (a), (b) and (c)(2) You can read the whole Act at this link HERE Also, for a good understanding of how all of this should work read HERE. The NTIA stance towards CPNI (customer proprietary network information) can be found HERE in a letter to the FCC. And then there is the SBC marketing department...which is the real reason. by RadioDoc edited by lev If you are interested in learning about Remote Terminals (RTs, CEVs, etc.) check out the Remote Terminal FAQ. A picture of one in the process of installation is here. Some remote terminals are in a Controlled Environmental Vault, which is an underground sealed room holding the equipment. The only part you'll see is the green entry and vent hatch, which is also usually landscaped and blends right in with it's surroundings. See »www.gsuc.net/index10.html for a short description with pictures of one being installed. If you are serviced via RT you can get any Sync speed package SBC offers. RT's are owned and serviced by the telco(ILEC) part of SBC not ASI(CLEC) this thread explains this. Here is another RT thread with pictures. An excellent explanation of why some remote terminals are and some aren't DSL-capable is from that same thread at /forum/remark,3418172~mode=flat#3438234. Finally, Lightspeed has, uh, seen the light! »SBC installed a big new curbside box in the neighborhood... by wayjac edited by lev This content was copied from a post by nunya on the AT&T subsidiaries that make AT&T Yahoo! DSL work. I posted on this a few months back, but it was an explanation of the DSL "trinity" at ATT. I'll give a basic rundown again. ATTIS - ATT Internet Services (formerly SBCIS). Technically speaking these people should be your SPOC (single point of contact) on DSL issues. When you call support, you usually get a contractor (who may or may not speak the Kings English, or be in the U.S. for that matter). Most people never make it past CSI (customer self install) or Tier 1. A second layer of support (There are more, but we need to keep this relatively simple) is also provided by ATTIS. They have their own Tier 1 and Tier 2 and NOC service (confused yet?) In the former Ameritech Region, ATTIS (content by Yahoo) is responsible for your DSL service. They operate all of the backbone that gets you from the Redback router (at your local POP) to the internet. They provide internet connectivity, authentication, passwords, usernames, ip addresses, blah, blah, and blah. ASI - "Advanced" Solutions Inc. - Another AT&T subsidiary, organizes and oversees the connection from your home to the DSLAM and Redback router. The DSL installation forces are also included in ASI. ASI is responsible for providing circuits for ATTIS DSL and other DSL ISP's. ASI works closely with the third party, the ILEC (AT&T Illinois, AT&T Ohio, AT&T Michigan, AT&T Wisconsin, and AT&T Indiana). ILEC -Incumbent Local Exchange Carrier - Does the physical wiring in the central office and outside plant up to your premise. Technically, you should never have to talk to ASI or the ILEC during your DSL life unless there is a major issue. While technically the ISP (ATTIS) should be the S.P.O.C., quite often people end up dealing with ASI or the ILEC. The three units are separated due to regulation. This could end soon and the most likely re-union would be ASI and the ILEC, while the ISP would remain separated. This content was copied from a post by lev about who you talk to for technical support When you call 877-722-3755 and ask for technical support (other than customer self-install), you start out talking to Sykes or Convergys. They're there to make sure that your problem isn't due to something that's easy to solve by a script. Creativity is not encouraged at those companies. If you legitimately have a problem, there is very little they can do. Since Convergys and Sykes get paid primarily based on their call volume, they're hard on employees that don't turn over calls quickly. And guess what takes a rep a lot of time? You guessed it. Going all the way through the script, and then having to sit on hold an additional 20 - 30 minutes to refer you to ATT-IS or ASI tier 2. The solution? You're here on this site. Post your problem in this forum. IF NECESSARY, someone will point you to post the problem in the ATT Direct area. Don't post it there first, primarily because they've had two (this last week, one) person doing the work that was too much for four a few months ago. Tier one has its uses. Really, they do. But don't expect them to be able to fix, or even know about major problems. ATT-IS and ASI don't keep them current with information. by wayjac Everything you could possibly want to know about the technical whys and becauses is in the official definition, RFC 2516. However, here's the meat of this protocol and how it works: PPP PPP is an acronym for Point to Point Protocol. PPP is a member of the TCP/IP suite of network protocols. PPP is an extension to TCP/IP that adds user ID and password security. LCP is an acronym for Link Control Protocol and is a component of PPP. IPCP is an acronym for IP Control Protocol and is a component of PPP. PPPoE PPPoE is an acronym for Point to Point Protocol over Ethernet PPP, which was designed for serial communications, has been combined with Ethernet, and is appropriately known as PPPoE. PPP was designed to do things that were either impossible or unnecessary with Ethernet, users are often confused as to why one would want to use PPP over Ethernet at all. The use of PPPoE, which allows ISP's and CLEC's to secure communications through the use of user ID's and passwords and have the ability to measure/monitor the traffic of each user. Some DSL providers don't use PPPoE. PPPoE is a standard, and you may think that all PPPoE clients are equal. As in most things, some are more equal than others. The standard requires that PPPoE clients add an additional header to the beginning of each TCP/IP packet. This may cause the packet to become larger than the maximum allowable size. Some PPPoE client solutions handle this transparently, but some require you to modify the TCP/IP settings on all of the client computers on the LAN. Some modem's show some of the exchange other's show it all. There are RFC's for all mentioned Protocols refer to them to understand the information given. Below is a very detailed PPPoE exchange from start to finish: 00:00:23 POE: Sending PADI packet for PPP 00:00:23 POE: ENET Received PADO packet 00:00:23 POE: Sending PADR packet for PPP 00:00:23 POE: ENET Received PADS packet 00:00:23 POE: PPP Session ID = 0x1e12 is open 00:00:23 PPP: LCP We processed peer's requests: 00:00:23 PPP: max-recv-unit, mru 1492 (ack) 00:00:23 PPP: auth-type, auth (ack) 00:00:23 PPP: magic-num, 0x41d8799 (ack) 00:00:23 PPP: LCP Peer acknowledged our requests: 00:00:23 PPP: mru 1492 00:00:23 PPP: magic-num, 0xea27fe02 00:00:23 PPP: LCP We processed peer's requests: 00:00:23 PPP: max-recv-unit, mru 1492 (ack) 00:00:23 PPP: auth-type, auth (ack) 00:00:23 PPP: magic-num, 0xdeeb76cb (ack) 00:00:23 PPP: LCP Peer acknowledged our requests: 00:00:23 PPP: mru 1492 00:00:23 PPP: magic-num, 0xb9aebba6 00:00:23 PPP: PAP authentication success reported by peer 00:00:23 IPCP: Resettings options: ip (0.0.0.0), dns (0.0.0.0, 0.0.0.0) 00:00:23 IPCP: IPCP We acknowledged peer's requests: 00:00:23 IPCP: IP-Addr: peer-addr 216.117.66.1 00:00:23 IPCP: Peer Nak'd our addr (0.0.0.0) with addr (216.117.108.2); accepting peer's address preference 00:00:23 IPCP: Peer Nak'd our primary DNS (0.0.0.0) with (209.172.192.2); accepting peer's primary DNS address 00:00:23 IPCP: Peer Nak'd our secondary DNS (0.0.0.0) with (209.172.194.2); accepting peer's secondary DNS address 00:00:23 IPCP: IPCP Peer acknowledged our request: 00:00:23 IPCP: IP-Addr: local-addr 216.117.108.2 00:00:23 IPCP: Primary DNS: 209.172.192.2 00:00:23 IPCP: Secondary DNS: 209.172.194.2 00:00:23 IPCP: IP up, local 216.117.108.2, remote 216.117.66.1 This is a breakdown/explaination of a 5100b modem log posted by core network with edits it may give some insight to what is going on with the PPPoE exchange. NOTE: The 5100b modem log has a time delay for its entries. 2004/05/06 00:56:13 GMT E |System |=========== SYSTEM UP =========== 2004/05/06 00:56:13 GMT E |System |Current Mode: PPP on the modem (Public IP for LAN device) 2004/05/06 00:56:14 GMT E |DSL |DataPump Version 01.01.00.00 2004/05/06 00:56:14 GMT E |DSL |State: WAITING Modem has finished booting up 2004/05/06 00:56:15 GMT E |Ethernet |Link 1 Up - 100Base-TX Full Duplex Ethernet port is up 2004/05/06 00:56:16 GMT E |DSL |State: INITIALIZING 2004/05/06 00:56:28 GMT E |DSL |State: WAITING 2004/05/06 00:56:30 GMT E |DSL |State: INITIALIZING Modem is doing its Synchronization routine 2004/05/06 00:56:43 GMT E |DSL |HYBRID 2 This line can also end with HYBRID 3 2004/05/06 00:56:38 GMT E |DSL |Link up 1 US 160 DS 1536 (FAST:G.dmt) you know what this step is - your negotiated speed with the DSLAM port itself. 00:56:38 GMT=Time Stamp GMTin EST this would be 7:56:38pm/19:56:38 E |DSL=Section of modem making this entry Link up 1=Number of times modem has achieved synchronization a reboot will reset counter to 0 US 160 DS 1536=Sync rates (FAST:G.dmt)=Fastpath/Interleaved Explanation (FAST:G.dmt)=Encoding standard TI.413 is also used 2004/05/06 00:56:46 GMT E |PPPoE |tx PADI, id: 0000, ac: (NULL), sn: (NULL) Here your modem/router is trying to initiate a PPPOE session. The PADI is the first step in a 4-way handshake that the PPPOE protocol uses for session establishment. tx=Transmit 2004/05/06 00:56:46 GMT E |PPPoE |rx AC Name: 62031020042800-ipltinho03w This is step number 2. The name here that you are seeing is actually the name of the ASI RedBack router. This is because your PPPOE session is actually being handled by ASI, not the ISP. The name of this packet being sent back to you is PADO. ASIs router is actually offering to act as the PPPOE server. This router is usually transparent to the end user, but it can be seen when looking at the output from a PPPoE client. It's also visible in the system properties section of the Enternet 300 software client. rx=Recieve AC Name=Access Concentrator/RedBack router 2004/05/06 00:56:46 GMT E |PPPoE |tx PADR, id: 0000, ac: (NULL), sn: 3. In this step your router/PPPOE client has received the PADO packet from ASIs router, and is requesting sending a PADR to establish a PPPOE session with this router, much the same way a NIC card will establish an IP address from a DHCP server. 2004/05/06 00:56:46 GMT E |PPPoE |rx PADS id: 2BAF After you send the PADR packet, then the last step of the PPPOE 4-way handshake is complete, the PADS packet that is sent back to you. Think of this as the PAD Standby while I authenticate your session with the ISP. This is because the ASI router does your ppp authentication for you this is known as proxy ppp. You have already sent me your login and password info by step 3. id: 2BAF=This is the Session ID 2004/05/06 00:56:46 GMT E |PPP |LCP neg PAP The beginning of the ppp authentication process LCP=Link Control Protocol PAP=Password Authentication Protocol 2004/05/06 00:56:46 GMT E |PPP |LCP up You login credentials have been received and approved. You are halfway there at this point. 2004/05/06 00:56:48 GMT E |PPP |IPCP nak option: 3 The NCP phase of ppp, or Network Control Protocol, has begun. Right now you are being told to standby for DNS info coming your way. 2004/05/06 00:56:48 GMT E |PPP |IPCP nak option: 129 This is an acknowledgement of the first DNS server address you have received. 2004/05/06 00:56:48 GMT E |PPP |IPCP nak option: 131 This is an acknowledgement of the second DNS server address you have received. 2004/05/06 00:56:48 GMT E |PPP |IPCP up ip: 68.251.190.179, gw: 68.251.191.254, dns: 67.36.128.26, 206.141.192.60 This should be obvious. Your IP address,gateway and DNS info have now been assigned. 2004/05/06 00:56:57 GMT E |SNTP Client |Updated time from Primary server 132.163.4.103 A timestamp of when your session began. SNTP Client=Simple Network Time Protocol Two public time servers are used to set and update the modems clock From this point on it should be happy trails - you are free to roam the beautiful world we call "internet". Have fun and watch out for infections. 2004/03/18 10:57:25 GMT E |DSL |Link Down 2004/03/18 10:57:25 GMT E |DSL |State: WAITING 2004/03/18 10:57:31 GMT E |DSL |State: INITIALIZING 2004/03/18 10:57:39 GMT E |DSL |HYBRID 2 2004/03/18 10:57:39 GMT E |DSL |Link up 8 US 160 DS 1536 (FAST:G.dmt) 2004/03/18 20:15:41 GMT E |DSL |Link Down 2004/03/18 20:15:41 GMT E |DSL |State: WAITING 2004/03/18 20:15:44 GMT E |DSL |State: INITIALIZING 2004/03/18 20:15:52 GMT E |DSL |HYBRID 2 2004/03/18 20:15:52 GMT E |DSL |Link up 9 US 160 DS 1536 (FAST:G.dmt) 2004/03/18 20:18:00 GMT E |DSL |Link Down 2004/03/18 20:18:00 GMT E |DSL |State: WAITING 2004/03/18 20:18:07 GMT E |DSL |State: INITIALIZING 2004/03/18 20:18:15 GMT E |DSL |HYBRID 2 2004/03/18 20:18:15 GMT E |DSL |Link up 10 US 160 DS 1536 (FAST:G.dmt) The section above shows sync loss three times but the PPP session (IP) does not go down 2004/03/21 09:31:17 GMT E |PPP |Max echo misses 2004/03/21 09:31:17 GMT E |PPP |LCP down 2004/03/21 09:31:17 GMT E |PPP |IPCP down 2004/03/21 09:31:35 GMT E |PPPoE |tx PADT, id:28AA, ac:(NULL), sn:(NULL) The section above shows the PPP session being Terminated by the modem in response to the Max echo misses by sending a PADT , the PPP session can also be terminated by the Access Concentrator/RedBack router if it sends a PADT or a loss of sync by wayjac edited by lev provided the content for this FAQ in this Thread ATM is used from the modem to the DSLAM, but the authentication and control, which is PPPoE, is encapsulated within the ATM cells. C. O. based DSLAM's are DS3 sized pipes that traverse across the ATM cloud, and eventually will aggregate with other DS3 connections to an OC3 port that feeds a particular RedBack at ASI. The Remote Terminal, being field based, are Telco owned and operated. The circuits will eventually terminate on an ASI RedBack, but must first traverse the Telco network. This includes circuits originally terminating on an OCD device, which concentrates the circuits of several individual Remote Terminals , then terminates them as 1 logical pipe to the OC3 port of the ASI RedBack. If the ASI RedBack has say 10 PVC's terminating on it's OC3 ingress port, 7 may be individual DSLAM PVC's, and the 3 remaining may be Telco Remote Terminal PVC's, which in turn may contain several Remote Terminal circuits within each connection. Where ASI sends the traffic to SBCIS, or any other ISP, from that point depends on customer density aggregated PVC count on the ISP end, proper traffic balancing, or any number of other factors. Once in the ATM cloud, for example the distance between the Chicago POP or the Elmhurst POP, and any RedBack router within the respective POP's, is irrelevant. Remember the RedBack you see in a trace route to your IP address is the ISP RedBack, not the ASI RedBack. This is because it is the first point at which you are at Layer3. Look to your Enternet300 properties (or whatever PPPoE client you are using), and you will see another RedBack name/Access Concentrator/AC - that is actually the ASI RedBack...the PPPoE server. How about a picture A brief description for some of the acronyms is here: List of common acronyms from the Bellsouth FAQ. by wayjac edited by lev Following is a link to the Navas Group's website with a pretty good explanation of the problem and how it affects the performance of TCP. »cable-dsl.home.att.net/# ··· symmetry fredcy offers the following additional information and a possible cure for some folks: For Linux users the free "Wonder Shaper" script does an excellent job of shaping traffic to reduce the effects of big upload or download jobs. It's particularly useful when running a BitTorrent client that's doing uploads. See »lartc.org/wondershaper/. For my 3000/384 ADSL line I found that setting wshaper to 2000/250 works well, for example. by rolande edited by lev The quoted "speed" is the sync rate between your modem and the DSLAM or RT port. This reflects the raw bitrate upon which everything else flows. Your data is encapsulated into Ethernet frames which are then encapsulated into TCP/IP packets which are then encapsulated into ATM cells which then are transmitted down the wire. said by salahx:at »SBC DSL bait and swtich is illegal...1500=750? : Actually the majority of the overhead isn't because of TCP/IP, its because of ATM (53 octet cells - 48 octet payload, 5 octet header). The PPPoE overhead is 8 octets per MTU (normally 1500 - this size of 1 Ethernet frame) The IP overhead is only 20 octets per MTU - PPPoE (normally 1492) and the TCP overhead is only 24 octets per MSS (normally 1452) So thus: PPPoE overhead is 0.53 % IP overhead is 1.30 % TCP overhead is 1.65 % ATM overhead, on the other hand is, 9.4 % For total DSL delivery overhead of about 12.9 %, or 87.1% efficient. Then there is the Ethernet overhead: Ethernet overhead bytes: 12 gap + 8 preamble + 14 header + 4 trailer = 38 bytes/packet w/o 802.1q 12 gap + 8 preamble + 18 header + 4 trailer = 42 bytes/packet with 802.1q Ethernet Payload data rates are thus: 1500/(38+1500) = 97.5293 % w/o 802.1q tags 1500/(42+1500) = 97.2763 % with 802.1q tags Best case we're out another 2.5 percent just for using Ethernet...or 97.5% of 87.1%, for a final "efficiency" of 84.9%. So therefore on a "perfect" 1536/384 line the max payload "speed" would be a tad over 1,300/326 kbps. A 3008/512 (sync) line delivers about 2554/435 (payload) and a 6016/768 (sync) line 5108/652 (payload). Furthermore, ATM has services on top of it like IP does with TCP and UDP, introducing more overhead, for example, bridging Ethernet over ATM. Also, many of the data protocols within IP/TCP/UDP have even MORE overhead, depending on the protocol. Every DSL provider does it the same way, the only way for a real improvement is to drop ATM, which isn't going to happen, as it is far too useful to drop, even with that kind of overhead Minor edits for clarity & additional info by RadioDoc by RadioDoc Q: OK, so exactly what is a DSL Filter and why is it needed and how do they work? A:Riss Centaur's 50-cent DSL Filter Primer: Way back in the dim reaches of time, the good old Bell System (Alexander Graham Bell, The Telephone Company, Ma Bell, Western Electric, Before Divestiture, the Big Breakup, whatever....) came up with the standard for the VOICE BAND or POTS (Plain Old Telephone Service) that all telephones should handle throughout the system. It was decided that telephones would need to carry signals from about 300Hz (cycles) through to about 3,300Hz. This is the range that is needed for intelligible speech. (In truth, the phone line band is 0 to 4,000Hz, just to be safe and carry touch tones.) This is also why music, with its higher frequencies (up to 20,000Hz) sounds lousy over a phone - because you lose the high frequencies. Ok, so voice is in the band of 300-3,300Hz. Where is DSL? ADSL uses the frequencies from about 26,000 (26K) to 1,100,000 (1,100K) Hz. These much higher frequencies are outside of human hearing, but when shoved down the phone wires can carry data to the modems! So when it is all put together it looks like this: We have to keep the Green and the Blue out of the POTS equipment and keep the POTS equipment from messing up that set of high frequency signals. So now the big moment! The dilemma is this: Most phones or any other type of equipment that connects to the phone line (answering machine, fax, Tivo, cable box, alarm system) were only designed to deal with Voice Band signals! If you suddenly start feeding DSL signals into them you can wind up with all sorts of problems - Noise in the voice band, Shorting out of the DSL signal, Crosstalk of noise into the DSL signal, and god only knows what else! Can you see the light come on? Bright Idea: Keep the DSL signal out of the equipment that does not know what to do with it. Install "The Filter"! The Filter - AKA a Microfilter, POTS filter, DSL filter, Splitter, Inline filter, plastic box with wires on it... you get the idea. Now, a filter works by filtering out the high frequency DSL signals, it does not let them pass through it, it only allows the low frequency voice signals to pass. Thats why in the electronic business this type of circuit is know as a Low Pass filter. It is a fairly simple idea, and a fairly simple circuit to implement with coils and capacitors and resistors. They are passive devices, they usually have no active electronics in them like transistors or IC's, and need no outside power. The filter, in addition to keeping the high frequency DSL signals out of equipment that does not know what to do with it, also keeps that equipment from loading down or shorting out the DSL signal itself! The two ways of filtering are: Install an inline filter on each phone and other device where it plugs into the wall - or - install a big filter at the main junction (NID) where the phone line comes in and it will handle filtering everything behind it. (And you have to run a home run wire from there to your DSL modem so it has an unfiltered signal.) Ok, what is the difference between a Filter and a Splitter eh? Now if you have a newer type filter that has 2 jacks, one marked VOICE and one marked DSL, you could think of it as a splitter. In truth all the DSL jack is is an extra connection off the plug fed straight through from the wall, so you have an extra jack to plug your DSL modem into. Also, filters installed at the NID are referred to as splitters, as they filter out the DSL signals from the wires split off of the main phone line. But Mr. Centaur, how do you install these things? I'm glad you asked. Lots has already been written on this subject. And NID Splitter installation: And as for where to get this little marvels, well, the microfilters usually come in your install kit, or can be had at electronic stores or online. POTS Splitters for installation at the NID can usually be had on Ebay in the $25-$30 range on good days. Now, arent you glad you asked? ;) Here are images of the guts of an NID Pots Splitter and a plug in Micro Filter (provided by d_l ) Corning/Seicor SPS-H50-SR1 Excelsus Z-330TJA Z-Blocker You will NOT get a reply if you click "Got Feedback" as that is for suggested changes to the FAQ only. If you have additional questions, visit the AT&T DSL Forum. Feedback received on this FAQ entry:
by Riss_Centaur edited by lev This information can be found in this FAQ: »:/faq/6734 Feedback received on this FAQ entry:
by Riss_Centaur edited by lev Many things can raise (or lower) the noise environment your DSL circuit works in, but since the attenuation is usually constant, the most important parameter--noise margin--can be affected by these changes. Drop too low and you lose sync. Noise margin is related to the bit loading, and is also affected by the DSLAM/RT port output signal power (download) or your modem's output signal power (upload). A forum discussion about it can be found here: »What causes SN Margin to change? |