My point was that there is no conspiracy here (I did read the article).
Besides, static IP addresses is a luxury from the old days. VPN technology should keep up with the world anno 2001. If dynamic IP addressing is so VPN hostile how comes that many VPN solutions work fine with traveling persons connecting to the company VPN through whatever local IP service they can use when traveling?
Adding the authentication and session concepts of PPPoE does not make for a more insecure network than static IPs and no authentication and no session context provides for currently. If anything it improves security.
PPPoE on a cloud of ATM connections is a great way for third party providers to use the phone companys lines. Another approach would be the basic ATM VCs - but that would require big investments in network infrastructure for both the telco and the provider.
Besides, there would be a number of problems and flights to be fought over service priorities.
If one ATM VC at 100% line capacity is used that connects to a pool of PPPoE server the users equioment can chose the correct PPPoE server and log in there. There is even good stuff for VPN solutions to be found here..
I dont know the details of SBCs plan, so the above (and all of this discussion) is based on facts and knowledge.
DOCISS has a bacground in cable modems.
DSL comes from the phone company.
The phone company is used to ATM and has invested incredible amounts of money in ATM infrastructure.
Using ATM for customer access seemed as a good idea.
In order to get a cheap and stable system they chose to use some standardized protocols on top of the ATM layer. One of them are PPP over ethernet carried over AAL5 PDUs on the DSL link.
This allowed them to use the existing equipment and management sytems that were all PPP oriented.
By using PPP authentication the telco does not have to install any additional hardware to identify the user when connecting. Keep in mind that the Telco might want to have authentication running from the first day in order to provide you with differentiated services at a later stage (QOS, voice etc.).
But basicly I think the authentication arises from the fact that telcos when DSL started to roll out thought they could bill the customer pr. minute or based on traffic. Now its all flat rate in order to be competitive however, I know that this reduces their need for the user to authenticate. But since there can be a multitude of reasons to do so in the future it is understandable that the telco might want you to authenticate with their service (lets forget about the legal issues in some contries for a moment in order not to stur up any conspiracy focused people).
PPPoE allows an ISP to sell one DSL line to one house, and then sell a number of internet connections on this link. This can be used for building complexes, dormatories, buisness centres etc.
The free selection of ISP based on PPPoE is very simple for the telco or provider to implement since PPPoE has the capability of selecting between multiple PPPoE servers. When the telco provides one physical link to the customer the different ISPs will need to have an authentication method for the user since only the telco has the linecard information (that anyway would require to much work for them to carry along their net and infrastructure).
In addition an ATM based third party solution would require a bigger investment for the little ISP player. PPPoE can be set up with an big IP connection and an access concentrator alone.
I dont think the PPPoE way will keep the small players of the field - even less that it is designed to do so.
So using PPPoE _is _ a clever solution. Of course a part of the bandwith are used by the protocol (6 bytes pr. PPP frame of 1500 bytes length)
And the multiprotocol over ATM encapsulation needed to get any form of IP running over the ATM links consumes 8 bytes plus ethernet source and destination addresses (used for IPoA as well).
I think that not PPPoE but the ATM idea of DSL links should be questioned. Once ATM for the links has been deployed PPPoE is a great solution when all aspects are considered. Of course a tech savy user would prefer "pure IP" over something very transparent, but what would that transparent something be? ATM solves a lot of problems and is a clever choice for DSL.
Besides ATM opens up for a lot of interesting features in the future such as voice over the DSL link in addition with (IP)data. ATM makes leveraging of existing technologies and standards possible - not only for the telco/ISP but also for the customer.
I think we can do without a couple of bytes here and there considering the overall picture.
Besides, at the speeds of ADSL, SDSL and SHDSL the lost bandwidth is not noticable.
My other idea of the telco selling ATM VCs to third parties was just a way of describing a way for the telco to bypass the PPPoE part. I am not a telco nor an ISP so I am not biased in their favor - I make my living by developing DSL access products.
"the average user needs authetication because of what?"
The ISP needs it in order for the book keeping of traffic, accounts and in some contrys security.
In a DSL setup the ISP has no other mean of matching a user to a connection.
You need authentication for the ISP to be able only to give service to their registered customers. Of course they could develop some smart way of matching up the incoming ATM connections from the access concentrators by interface caracteristics or similar. But the simpler (and thus cheeper for the customer) solution is authentication. Be it your MAC address for the IPoA/DHCP approach or PPP username for the PPPoX methods.
Clear enough?
1)
Why is PPPoE diffenrent from PPPoA or IPoA in this respect? It is just another protocol. If other ISPs could buy raw ATM channels from their CO to the customer it would be up to ISP #1 to chose which protocol to use. Will SBC allow PPPoE only third party ISPs?
2)
Capped - in what sense do you mean?
3)
No, but he(or she) might want a cable from one operator and IP service from another in cases where proper cable coverage is only available from one cable operator. This might be settled between the cable operator and the ISP, but it might be that the end user can do the shopping.
However for DSL connections this can only be done the correct way if the ISP rents an ATM connection to the customer. But wait a moment - is this a discussion of PPPoE and its crappy-level? I dont think so...
4)
In a free contry a company is free to put together the product it markets. If the company wants to differentiate between two market segments it is free to do so. The same fredom applies to you - you can buy the buisness product if you want the better performance or the higher geek-factor.
The only way the ISPs can provice low priced residential IP service is by optimizing their network and operational infrastructure. PPP-over-what-ever is a great solution for doing that.
Please proff that this is what the Telcos are planning to do. It sounds like just another adolescent conspiracy theory,
I still do not get the point of static/dynamic IP assignment vs. high/low end level of service.
And it might be that line providers are better of selling ATM connections to third party providers. What would SBC gain from opening up for third parties PPPoE connections?
Could you guys perhaps use an hour or so reading up on PPP, ATM, DSL technologies and the different issues an ISP will have to tackle in order to provide you with your beloved bandwidth?
PPPoE isnt bad, PPPoE doesnt prohibit static IP. DHCP is not comparable with PPPoE.
The comparison would have to be between PPPoE (over AAL5) (or PPP over AAL5) and IP over AAL5 the RFC1483 way.
Compare the tree and decide what you would implement if you were to make money in that buisness and had to plan for 100.000+ customers.
The only real viable solution is PPPoE over AAL5.
Slashdot Mirror
Anyone up for doing a receiver only using a standard PC and what ever lies on an ordinary desktop (paper, paperclips, eraser etc.) ??
My point was that there is no conspiracy here (I did read the article). Besides, static IP addresses is a luxury from the old days. VPN technology should keep up with the world anno 2001. If dynamic IP addressing is so VPN hostile how comes that many VPN solutions work fine with traveling persons connecting to the company VPN through whatever local IP service they can use when traveling? Adding the authentication and session concepts of PPPoE does not make for a more insecure network than static IPs and no authentication and no session context provides for currently. If anything it improves security. PPPoE on a cloud of ATM connections is a great way for third party providers to use the phone companys lines. Another approach would be the basic ATM VCs - but that would require big investments in network infrastructure for both the telco and the provider. Besides, there would be a number of problems and flights to be fought over service priorities. If one ATM VC at 100% line capacity is used that connects to a pool of PPPoE server the users equioment can chose the correct PPPoE server and log in there. There is even good stuff for VPN solutions to be found here.. I dont know the details of SBCs plan, so the above (and all of this discussion) is based on facts and knowledge.
DOCISS has a bacground in cable modems. DSL comes from the phone company. The phone company is used to ATM and has invested incredible amounts of money in ATM infrastructure. Using ATM for customer access seemed as a good idea. In order to get a cheap and stable system they chose to use some standardized protocols on top of the ATM layer. One of them are PPP over ethernet carried over AAL5 PDUs on the DSL link. This allowed them to use the existing equipment and management sytems that were all PPP oriented. By using PPP authentication the telco does not have to install any additional hardware to identify the user when connecting. Keep in mind that the Telco might want to have authentication running from the first day in order to provide you with differentiated services at a later stage (QOS, voice etc.). But basicly I think the authentication arises from the fact that telcos when DSL started to roll out thought they could bill the customer pr. minute or based on traffic. Now its all flat rate in order to be competitive however, I know that this reduces their need for the user to authenticate. But since there can be a multitude of reasons to do so in the future it is understandable that the telco might want you to authenticate with their service (lets forget about the legal issues in some contries for a moment in order not to stur up any conspiracy focused people). PPPoE allows an ISP to sell one DSL line to one house, and then sell a number of internet connections on this link. This can be used for building complexes, dormatories, buisness centres etc. The free selection of ISP based on PPPoE is very simple for the telco or provider to implement since PPPoE has the capability of selecting between multiple PPPoE servers. When the telco provides one physical link to the customer the different ISPs will need to have an authentication method for the user since only the telco has the linecard information (that anyway would require to much work for them to carry along their net and infrastructure). In addition an ATM based third party solution would require a bigger investment for the little ISP player. PPPoE can be set up with an big IP connection and an access concentrator alone. I dont think the PPPoE way will keep the small players of the field - even less that it is designed to do so. So using PPPoE _is _ a clever solution. Of course a part of the bandwith are used by the protocol (6 bytes pr. PPP frame of 1500 bytes length) And the multiprotocol over ATM encapsulation needed to get any form of IP running over the ATM links consumes 8 bytes plus ethernet source and destination addresses (used for IPoA as well). I think that not PPPoE but the ATM idea of DSL links should be questioned. Once ATM for the links has been deployed PPPoE is a great solution when all aspects are considered. Of course a tech savy user would prefer "pure IP" over something very transparent, but what would that transparent something be? ATM solves a lot of problems and is a clever choice for DSL. Besides ATM opens up for a lot of interesting features in the future such as voice over the DSL link in addition with (IP)data. ATM makes leveraging of existing technologies and standards possible - not only for the telco/ISP but also for the customer. I think we can do without a couple of bytes here and there considering the overall picture. Besides, at the speeds of ADSL, SDSL and SHDSL the lost bandwidth is not noticable. My other idea of the telco selling ATM VCs to third parties was just a way of describing a way for the telco to bypass the PPPoE part. I am not a telco nor an ISP so I am not biased in their favor - I make my living by developing DSL access products.
"the average user needs authetication because of what?" The ISP needs it in order for the book keeping of traffic, accounts and in some contrys security. In a DSL setup the ISP has no other mean of matching a user to a connection. You need authentication for the ISP to be able only to give service to their registered customers. Of course they could develop some smart way of matching up the incoming ATM connections from the access concentrators by interface caracteristics or similar. But the simpler (and thus cheeper for the customer) solution is authentication. Be it your MAC address for the IPoA/DHCP approach or PPP username for the PPPoX methods. Clear enough? 1) Why is PPPoE diffenrent from PPPoA or IPoA in this respect? It is just another protocol. If other ISPs could buy raw ATM channels from their CO to the customer it would be up to ISP #1 to chose which protocol to use. Will SBC allow PPPoE only third party ISPs? 2) Capped - in what sense do you mean? 3) No, but he(or she) might want a cable from one operator and IP service from another in cases where proper cable coverage is only available from one cable operator. This might be settled between the cable operator and the ISP, but it might be that the end user can do the shopping. However for DSL connections this can only be done the correct way if the ISP rents an ATM connection to the customer. But wait a moment - is this a discussion of PPPoE and its crappy-level? I dont think so... 4) In a free contry a company is free to put together the product it markets. If the company wants to differentiate between two market segments it is free to do so. The same fredom applies to you - you can buy the buisness product if you want the better performance or the higher geek-factor. The only way the ISPs can provice low priced residential IP service is by optimizing their network and operational infrastructure. PPP-over-what-ever is a great solution for doing that.
Please proff that this is what the Telcos are planning to do. It sounds like just another adolescent conspiracy theory, I still do not get the point of static/dynamic IP assignment vs. high/low end level of service. And it might be that line providers are better of selling ATM connections to third party providers. What would SBC gain from opening up for third parties PPPoE connections?
Could you guys perhaps use an hour or so reading up on PPP, ATM, DSL technologies and the different issues an ISP will have to tackle in order to provide you with your beloved bandwidth? PPPoE isnt bad, PPPoE doesnt prohibit static IP. DHCP is not comparable with PPPoE. The comparison would have to be between PPPoE (over AAL5) (or PPP over AAL5) and IP over AAL5 the RFC1483 way. Compare the tree and decide what you would implement if you were to make money in that buisness and had to plan for 100.000+ customers. The only real viable solution is PPPoE over AAL5.
Oh wait, wasnt that already done? How hard can it be to implement such a simple protocol? Even Microsofts SQL server is fairly secure.