• How does it work on the technical level when you change ISP ?

    From Spiros Bousbouras@21:1/5 to All on Sun Jul 25 09:42:12 2021
    Let's say that your ISP is company A. Then the packets leaving your router
    pass through servers belonging to A. At some point you change your ISP to company B. Then the packets will go through the servers belonging to B. Which device(s) decide where to send the packets leaving your router and how does it/they know to which provider to send them ? I understand that there are probably many different set ups , I'm just asking for an idea how it might work.

    --
    vlaho.ninja/prog

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andy Burns@21:1/5 to Spiros Bousbouras on Sun Jul 25 10:52:07 2021
    Spiros Bousbouras wrote:

    Let's say that your ISP is company A. Then the packets leaving your router pass through servers belonging to A. At some point you change your ISP to company B. Then the packets will go through the servers belonging to B.

    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or B,
    all traffic then travels through the tunnel.

    </shortversion>

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andy Burns@21:1/5 to Andy Burns on Sun Jul 25 10:57:32 2021
    Andy Burns wrote:

    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or B,
    all traffic then travels through the tunnel.

    For ADSL the wiring will need to be changed in the exchange (aka central office) to connect you to company B's equipment, for VDSL it's usually a logical change, by the line provider to let you reach company B, rather
    than a physical wiring change.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Spiros Bousbouras@21:1/5 to Andy Burns on Sun Jul 25 10:48:22 2021
    On Sun, 25 Jul 2021 10:52:07 +0100
    Andy Burns <usenet@andyburns.uk> wrote:
    Spiros Bousbouras wrote:

    Let's say that your ISP is company A. Then the packets leaving your router pass through servers belonging to A. At some point you change your ISP to company B. Then the packets will go through the servers belonging to B.

    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or B,
    all traffic then travels through the tunnel.

    How does my router know whether to communicate with the routers of A or B ?
    How does it know their IP addresses for example ?

    Regarding authentication , how does it know the right credentials ? For
    example en.wikipedia.org/wiki/Challenge-Handshake_Authentication_Protocol
    says
    CHAP requires that both the client and server know the plaintext of the
    secret, although it is never sent over the network.

    So I guess that my router and the router of the ISP must know some common secret. How does this get established ?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andy Burns@21:1/5 to Spiros Bousbouras on Sun Jul 25 15:31:18 2021
    Spiros Bousbouras wrote:
    On Sun, 25 Jul 2021 10:52:07 +0100
    Andy Burns <usenet@andyburns.uk> wrote:
    Spiros Bousbouras wrote:

    Let's say that your ISP is company A. Then the packets leaving your router >>> pass through servers belonging to A. At some point you change your ISP to >>> company B. Then the packets will go through the servers belonging to B.

    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or B,
    all traffic then travels through the tunnel.

    How does my router know whether to communicate with the routers of A or B ? How does it know their IP addresses for example ?

    depends, you're not giving many clues in terms of which continent you're
    on, or what connection type you have, these things do vary a bit.

    If it's ADSL, your router hasn't got much choice, the wires from your
    house lead straight to the equipment (DSLAM) for your ISP

    Regarding authentication , how does it know the right credentials ? For example en.wikipedia.org/wiki/Challenge-Handshake_Authentication_Protocol says
    CHAP requires that both the client and server know the plaintext of the
    secret, although it is never sent over the network.

    In this country, only BT doesn't require authentication, because they
    know which phone line you're on and use that instead

    So I guess that my router and the router of the ISP must know some common secret. How does this get established ?

    all other ISPs either require you to put a username and password in
    the router, or they pre-configure it before they send the router out to
    you, or use an automatic configuration system TR-069 which uses an
    identifier (e.g. serial number) of the router to setup the parameters
    for you.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Rich@21:1/5 to Spiros Bousbouras on Sun Jul 25 16:22:08 2021
    Spiros Bousbouras <spibou@gmail.com> wrote:
    On Sun, 25 Jul 2021 10:52:07 +0100
    Andy Burns <usenet@andyburns.uk> wrote:
    Spiros Bousbouras wrote:

    Let's say that your ISP is company A. Then the packets leaving
    your router pass through servers belonging to A. At some point
    you change your ISP to company B. Then the packets will go
    through the servers belonging to B.

    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or
    B, all traffic then travels through the tunnel.

    How does my router know whether to communicate with the routers of A
    or B ? How does it know their IP addresses for example ?

    You tell it via the "default route" you setup when you configure it.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Spiros Bousbouras@21:1/5 to Rich on Sun Jul 25 17:42:27 2021
    On Sun, 25 Jul 2021 16:22:08 -0000 (UTC)
    Rich <rich@example.invalid> wrote:
    Spiros Bousbouras <spibou@gmail.com> wrote:
    On Sun, 25 Jul 2021 10:52:07 +0100
    Andy Burns <usenet@andyburns.uk> wrote:

    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or
    B, all traffic then travels through the tunnel.

    How does my router know whether to communicate with the routers of A
    or B ? How does it know their IP addresses for example ?

    You tell it via the "default route" you setup when you configure it.

    I never did any such configuring on my current router. But , since it was provided by my ISP , it probably came preconfigured. But this leads to
    another question : what if someone enters a route not for an ISP they are subscribing to ? I'm guessing that some authentication will fail but if they somehow manage to crack the authentication process then they can get internet access without paying an ISP , right ? I was reading earlier en.wikipedia.org/wiki/TR-069 and www.pcworld.com/article/2463480/many-home-routers-supplied-by-isps-can-be-compromised-en-masse-researchers-say.html

    and they say that routers are crackable. The setting is different but if an outsider can crack a router then I'm guessing that the owner can too.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Spiros Bousbouras on Sun Jul 25 12:23:56 2021
    On 7/25/21 3:42 AM, Spiros Bousbouras wrote:
    Let's say that your ISP is company A. Then the packets leaving your
    router pass through servers belonging to A. At some point you change
    your ISP to company B. Then the packets will go through the servers
    belonging to B.

    Please clarify the timing that you're talking about when you say "at
    some point". Are you referring to a one time change - say ISP B is new
    and offers a sale to get you to switch to them, or are you referring to
    some special connection that uses both ISP A and ISP B /concurrently/ (depending on various configurations)?

    The former is trivial and usually involves someone; ISP A, ISP B, or
    common (telco) carrier C, making a change to ""re-wire your connection
    from A to B.

    The latter is possible, but usually is a much more complex configuration.

    There is also an additional varient where you have two completely
    independent connections, one to ISP A (xDSL) /and/ one to ISP B (DOCSIS).

    Which device(s) decide where to send the packets leaving your router
    and how does it/they know to which provider to send them ?

    It depends.

    I understand that there are probably many different set ups , I'm
    just asking for an idea how it might work.

    I hope the first scenario describes itself.

    The second scenario depends on the access technology, how it's
    configured, and what ISPs A and B support.

    The third scenario is decided by your router picking which ISP
    connection to use; A or B.

    Please ask clarifying questions.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Andy Burns on Sun Jul 25 12:28:36 2021
    On 7/25/21 3:57 AM, Andy Burns wrote:
    For ADSL the wiring will need to be changed in the exchange (aka central office) to connect you to company B's equipment, for VDSL it's usually a logical change, by the line provider to let you reach company B, rather
    than a physical wiring change.

    That may be the case where you are. But where I'm from there is no
    strict /need/ for any physical change.

    Where I'm from the DSLAMs are almost always owned by the common carrier telephone company (the Incumbent Local Exchange Carrier / ILEC). They
    would logically change the configuration for a given line's DSLAM port
    so that the connection changes from ISP A to ISP B. Quite like you're
    alluding to with VDSL.

    DOCSIS is a little bit different. Purportedly the DOCSIS /technology/
    supports multiple ISP's sharing a head end and utilizing a common
    outside plant Hybrid Fiber / Coax (HFC) network. However, I've never
    heard of this being done in practice. This works quite similar to a
    standard layer 2 switched Ethernet network. Each ISP simply looks like
    a different Ethernet device across the DOCSIS / HFC network.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Spiros Bousbouras@21:1/5 to Andy Burns on Sun Jul 25 18:30:56 2021
    On Sun, 25 Jul 2021 15:31:18 +0100
    Andy Burns <usenet@andyburns.uk> wrote:
    Spiros Bousbouras wrote:
    On Sun, 25 Jul 2021 10:52:07 +0100
    Andy Burns <usenet@andyburns.uk> wrote:
    For xDSL networks, your packets reach routers operated by A or B,
    because your router authenticates and creates a PPP tunnel to A or B,
    all traffic then travels through the tunnel.

    How does my router know whether to communicate with the routers of A or B ? How does it know their IP addresses for example ?

    depends, you're not giving many clues in terms of which continent you're
    on, or what connection type you have, these things do vary a bit.

    I'm asking for my general education so any reply for any location and any
    set up would be interesting to me.

    If it's ADSL, your router hasn't got much choice, the wires from your
    house lead straight to the equipment (DSLAM) for your ISP

    Regarding authentication , how does it know the right credentials ? For example en.wikipedia.org/wiki/Challenge-Handshake_Authentication_Protocol says
    CHAP requires that both the client and server know the plaintext of the
    secret, although it is never sent over the network.

    In this country, only BT doesn't require authentication, because they
    know which phone line you're on and use that instead

    Where "this country" means the U.K. So if it's an xDSL network then it is physically possible for the packets to travel from any home router to several ISPs. Who owns the routers which decide which ISP the packets go to ? Is
    it BT ?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Spiros Bousbouras on Sun Jul 25 13:00:11 2021
    On 7/25/21 12:30 PM, Spiros Bousbouras wrote:
    I'm asking for my general education so any reply for any location
    and any set up would be interesting to me.

    That ask actually makes it harder to answer. That's sort of like asking
    for a general description of how you get from New York to Los Angeles.
    There are many different modes of travel; plain, train, automobile,
    walking, boat, etc. Then there are many routes that could be used; east
    to west (~direct), or east to west (indirect around the world).

    Heck, there's even a chance that you could use a propulsion system to
    jump almost straight up, slightly south trajectory, such that when you
    came down the Earth would have rotated under you that you now land on
    the west coast. The southerly trajectory is to account for the
    latitudinal differences.

    You *REALLY* do need to refine your question /some/ to get more
    meaningful answers.

    So if it's an xDSL network then it is physically possible for the
    packets to travel from any home router to several ISPs.

    Yes, it is (physically and logically) possible for a home router to send packets to multiple ISPs /if/ the DSL network has been properly configured.

    I'm wording this as if it's an ATM DSL connection, but the same concepts
    exist with other types of DSL connections. Similar concepts also apply
    to other types of connections.

    Where I'm from, the ILEC owns the DSLAMs. So, ISP customers need their
    phone line to be connected to a DSLAM. This is /usually/ requested by
    the ISP, but I've heard of situations where the customer requests this directly. The DSLAM is connected to the ATM network (think cloud). The
    ISP is also connected to the ATM network. The ATM network owner
    provisions a PVC from the DSLAM to the ISP for the given customer. Thus
    when the ISP customer sends traffic, it goes over the DSL to the DSLAM,
    through the ATM network to the ISP, and the ISP sees the traffic coming
    in on the logical interface that represents the customer. The ISP then
    routes the traffic through their IP network out to the Internet.

    Aside: Depending on the size and / or complexity of the ISP's network,
    there may be other networking technologies used between their ATM router
    where DSL customer circuits terminate, through the ISP's core network,
    out to the Internet.

    There is no /technical/ reason why you can't have multiple PVCs from the
    same DSLAM port that represents the subscriber's phone line, each going
    to different ISPs. You could extend this to be two different PVCs to
    the same ISP. Presumably, each PVC would terminate in different
    locations. Though technically they could terminate in the same location
    as separate virtual circuits for different types of traffic.

    Who owns the routers which decide which ISP the packets go to ? Is
    it BT ?

    It depends.

    You could make an argument that it's the DSLAM / ATM network owner's
    equipment / configuration there of that decide which ISP the packets go to.

    You could also make an argument that it's the customer's equipment that
    picks the PVC (VPI & VCI) that traffic is sent down and that the DSLAM /
    ATM network is simply passing things according to the PVC information.

    Usually, there is only one PVC configured and the DSLAM / ATM network
    operator is who decides how to ""route things.

    Aside:
    - The term "switching" is used at the ATM level in lieu of "routing".
    - "Routing" is normally applied to IP packets at layer 3.
    - The term "cell" is used at the ATM level in lieu of "packet".
    - "Packets" usually refers to units of IP information at layer 3.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Spiros Bousbouras@21:1/5 to Grant Taylor on Sun Jul 25 18:35:51 2021
    On Sun, 25 Jul 2021 12:23:56 -0600
    Grant Taylor <gtaylor@tnetconsulting.net> wrote:
    On 7/25/21 3:42 AM, Spiros Bousbouras wrote:
    Let's say that your ISP is company A. Then the packets leaving your
    router pass through servers belonging to A. At some point you change
    your ISP to company B. Then the packets will go through the servers belonging to B.

    Please clarify the timing that you're talking about when you say "at
    some point". Are you referring to a one time change - say ISP B is new
    and offers a sale to get you to switch to them,

    Yes , this is the scenario I was thinking but , as I've said in a different post , I'm asking for my general education so hearing about other scenarios
    is fine.

    or are you referring to
    some special connection that uses both ISP A and ISP B /concurrently/ (depending on various configurations)?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Andy Burns on Sun Jul 25 12:40:36 2021
    On 7/25/21 8:31 AM, Andy Burns wrote:
    If it's ADSL, your router hasn't got much choice, the wires from your
    house lead straight to the equipment (DSLAM) for your ISP

    ADSL in particular, and xDSL in general as I understand it, supports
    common DSLAMs. (See my reply a few minutes ago for more details.)

    ATM based DSL supports multiple PVCs across the single physical
    connection. -- This would be called a Service Access Loop in ATM
    parlance. -- Thus ATM based DSL /technically/ /supports/ concurrent connections to multiple ISPs through the use of different PVCs.

    I have used ATM DSL modems that supported up to eight different PVCs.
    As in the modem itself established the PVC using the PVI and SVI
    provided by the router.

    In this country, only BT doesn't require authentication, because they
    know which phone line you're on and use that instead

    That's because there is association with the physical port on the DSLAM
    with the phone line that is connected. This pre-defined circuit makes
    it from the DSLAM through the provider's telco network to the router on
    the far side. The router has different logical interfaces that
    represent the physical DSLAM port. So there is a one-for-one mapping.
    As long as the association / mapping is maintained properly, you can
    easily know that logical port XYZ123 is phone number ABC-9876.

    all other ISPs either require you to put a username and password in the router, or they pre-configure it before they send the router out to you,

    That may be the case where you are. But where I'm from, all three of
    the ADSL ISPs (1 x ILEC and 2 x independent) used the one-for-one
    mapping described above. The PPPoE / DHCP / static configuration is
    actually (bridged / encapsulated) Ethernet frame riding across the DSL
    network -- which is more akin to point-to-point logical Ethernet links
    in this regard -- to the ISP who provides the counterpart for PPPoE /
    DHCP / static configuration.

    or use an automatic configuration system TR-069 which uses an identifier (e.g. serial number) of the router to setup the parameters for you.

    Things started with DHCP & static configuration and then progressed to
    PPPoE where I'm from. The PPPoE enabled providers to care less (or not
    at all) about the one-for-one mapping described above and instead rely
    on PPPoE's authentication capabilities. Thus the DSL became even more
    of a point-to-point Ethernet LAN. Or, more accurately point (ISP) to multi-point (DSL subscribers) as subscriber to subscriber still needed
    to pass through the ISP's head end, for both layer 2 and layer 3.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andy Burns@21:1/5 to Grant Taylor on Sun Jul 25 20:03:17 2021
    Grant Taylor wrote:

    On 7/25/21 8:31 AM, Andy Burns wrote:
    If it's ADSL, your router hasn't got much choice, the wires from your
    house lead straight to the equipment (DSLAM) for your ISP

    ADSL in particular, and xDSL in general as I understand it, supports
    common DSLAMs.  (See my reply a few minutes ago for more details.)

    For ADSL smaller ISPs here likely use BT's DSLAMs, larger ISPs have
    provided their own, that's changing as more move to VDSL and use common
    MSANs where the fibre backhaul reaches the street cabinet.

    ATM based DSL supports multiple PVCs across the single physical
    connection.

    But do any providers make use of that? Technically it's possible but
    here it's never used. It might have been nice to provision a home ISP
    and a work ISP on the same line using different VCI/VPI numbers, most
    providers used the same VCI/VPI nunbers, I can only think of one that
    even used different numbers, then decide on a split of bandwidth for
    each, but in practice most home ADSL users want more bandwidth than is available, so you either end up using home's bandwidht for work, or
    having a separate line.

      --  This would be called a Service Access Loop in ATM
    parlance.  --  Thus ATM based DSL /technically/ /supports/ concurrent connections to multiple ISPs through the use of different PVCs.

    But do they *use* that where you are either?

    I have used ATM DSL modems that supported up to eight different PVCs. As
    in the modem itself established the PVC using the PVI and SVI provided
    by the router.

    In this country, only BT doesn't require authentication, because they
    know which phone line you're on and use that instead

    That's because there is association with the physical port on the DSLAM
    with the phone line that is connected.  This pre-defined circuit makes
    it from the DSLAM through the provider's telco network to the router on
    the far side.  The router has different logical interfaces that
    represent the physical DSLAM port.  So there is a one-for-one mapping.
    As long as the association / mapping is maintained properly, you can
    easily know that logical port XYZ123 is phone number ABC-9876.

    all other ISPs either require you to put a username and password in
    the router, or they pre-configure it before they send the router out
    to you,

    That may be the case where you are.  But where I'm from, all three of
    the ADSL ISPs (1 x ILEC and 2 x independent) used the one-for-one
    mapping described above.

    Yep, but the only ISP that uses that here is the one who is BT, the
    owner of the kit in the exchanges (ok there's an arm's length separation between BT Openreach who own the kit and the wires, and BT Internet who
    are the ISP)

    The PPPoE / DHCP / static configuration is
    actually (bridged / encapsulated) Ethernet frame riding across the DSL network -- which is more akin to point-to-point logical Ethernet links
    in this regard -- to the ISP who provides the counterpart for PPPoE /
    DHCP / static configuration.

    I'm sure it varies around the planet, here ADSL uses ATM and PPPoE, VDSL
    uses PPPoE

    or use an automatic configuration system TR-069 which uses an
    identifier (e.g. serial number) of the router to setup the parameters
    for you.

    Things started with DHCP & static configuration and then progressed to
    PPPoE where I'm from.

    PPoE still uses authentication here.

    The PPPoE enabled providers to care less (or not
    at all) about the one-for-one mapping described above and instead rely
    on PPPoE's authentication capabilities.  Thus the DSL became even more
    of a point-to-point Ethernet LAN.  Or, more accurately point (ISP) to multi-point (DSL subscribers) as subscriber to subscriber still needed
    to pass through the ISP's head end, for both layer 2 and layer 3.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Andy Burns on Sun Jul 25 13:36:14 2021
    On 7/25/21 1:03 PM, Andy Burns wrote:
    For ADSL smaller ISPs here likely use BT's DSLAMs, larger ISPs have
    provided their own, that's changing as more move to VDSL and use common
    MSANs where the fibre backhaul reaches the street cabinet.

    I suspect the elephant in the room is the difference in population
    density where you are vs where I am (from).

    Where I'm from there is a lot of rural area where a single DSLAM may
    have trouble covering line length to connect to 32 houses. So 8 and 16
    port DSLAMs were common. In such sparse areas, each ISP deploying their
    own DSLAM would be *EXTREMELY* inefficient if not wasteful.

    But do any providers make use of that? Technically it's possible but
    here it's never used.

    It depends how you look at things.

    The ILEC in my home town was particularly obstinate, and well known in
    the country for being such. My understanding, based on reading that
    I've done, is that more cooperative ILECs / CLECs /would/ do multiple
    PVCs on xDSL connections. These LECs were usually found on the east &
    west costs. I was stuck in the middle of the U.S.A. where we seemed to
    be a decade behind in technology deployments.

    If you remove the xDSL aspect from the link and instead focus on ATM
    SALs, yes, such was done in my home town. Particularly with companies
    like ISPs connecting into the ATM network to receive PVCs from all of
    their customers. I'm told that such was also done for corporate /
    enterprise companies for PVCs out to branch offices. After all, there
    is exceedingly little difference between an ISP's connections to
    residential users and an Enterprise's connections to branch offices. At
    least form an ATM network operator's point of view.

    It might have been nice to provision a home ISP and a work ISP on
    the same line using different VCI/VPI numbers,

    I tilted at this windmill from the standpoint of my customers. I tried
    to get the phone company to provision a PVC from each of my customers
    locations to a common ISP and then between their locations. The idea
    was to divide the traffic such that site-to-site traffic went over one
    PVC and site-to-Internet traffic went over a different PVC.
    Unfortunately / sadly, I never got the cooperation from the ILEC. So
    instead, all site-to-site traffic went over the PVC to the ISP where it tromboned to other sites.

    most providers used the same VCI/VPI nunbers,

    VPI 8 and VCI 35 were quite common in the mid-west. I remember hearing
    of other common pairs for different regions of the U.S.A.

    The very nature of how ATM works means that all the customers could use
    the same VPI / VCI pair without conflicting with each other. :-)

    I can only think of one that even used different numbers, then decide
    on a split of bandwidth for each, but in practice most home ADSL
    users want more bandwidth than is available, so you either end up
    using home's bandwidht for work, or having a separate line.

    I had a number of customers that had an xDSL connection for the servers
    (on the Internet with static IPs) and a DOCSIS connection for general
    Internet surfing (access to the Internet via NAT). Then it simply
    became a configuration of which default gateway does this device use;
    .254 for DOCSIS / NAT or .253 for xDSL. (DHCP provided .254. Servers / specific clients were configured with .253 for special use cases.)
    There was usually some sort of fail over outgoing access to the Internet
    on each router in case it's local / preferred connection was down.

    But do they *use* that where you are either?

    I never heard about it being used on ADSL connections in my area. I did
    hear about it being used on ADSL connections on the coasts.

    Voice over DSL was an option on the coasts. My ILEC laughed at me when
    I asked about VoDSL in my area.

    Yep, but the only ISP that uses that here is the one who is BT, the
    owner of the kit in the exchanges (ok there's an arm's length separation between BT Openreach who own the kit and the wires, and BT Internet who
    are the ISP)

    We have a similar arms reach / legal business entities owned by the same corporate overlord, between the ILEC who owned the cable plant and
    DSLAMs and the and the ILEC with their ISP hat on.

    I saw cases where the same ISP would use more than one different method depending on the customer. Basic DHCP / static for some customers and
    PPPoE for other customers. I don't recall what the differentiation was.
    It may have been a multi-year migration. I did my best to avoid PPPoE because of the complications related to MTU / MSS.

    I'm sure it varies around the planet, here ADSL uses ATM and PPPoE, VDSL
    uses PPPoE

    ADSL started as Frame Relay, migrated to ATM, and then migrated to
    something more IP native. (I don't know the particulars about the last configuration as I changed job roles.)

    PPoE still uses authentication here.

    I'm not aware of PPPoE, or it's underlying PPP protocol, /not/ using /authentication/. Even if the credentials are null, they are still used
    by the underlying PPP protocol.

    Aside: Yes, "PPP protocol" is accurate because "PPP" is a proper name identifying specific technology. I could have said "underlying Point
    to Point Protocol" if I wanted to expand it.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Scott Dorsey@21:1/5 to spibou@gmail.com on Sun Jul 25 21:09:33 2021
    Spiros Bousbouras <spibou@gmail.com> wrote:
    Where "this country" means the U.K. So if it's an xDSL network then it is >physically possible for the packets to travel from any home router to several >ISPs. Who owns the routers which decide which ISP the packets go to ? Is
    it BT ?

    BT determines where the cables are connected up. In the case of DSL,
    the DSLAM can be connected to any one of a number of outgoing networks.
    In the case of other kinds of telco service, likely someone has to go
    out and punch down cross-connect wires from one place to another.

    But either it is a virtual or a physical wire from one point to another.
    It's not a cloud, it's a single line.
    --scott
    --
    "C'est un Nagra. C'est suisse, et tres, tres precis."

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Scott Dorsey@21:1/5 to gtaylor@tnetconsulting.net on Sun Jul 25 21:13:43 2021
    Grant Taylor <gtaylor@tnetconsulting.net> wrote:
    Where I'm from there is a lot of rural area where a single DSLAM may
    have trouble covering line length to connect to 32 houses. So 8 and 16
    port DSLAMs were common. In such sparse areas, each ISP deploying their
    own DSLAM would be *EXTREMELY* inefficient if not wasteful.

    Where I am, we have low population density, but only one DSL provider,
    namely the telco that owns the local loops. So that's who owns the
    DSLAM and when you get out onto the vast and dizzying internet, you are
    doing so from an IP address owned by them. This is typical of much of
    rural America.
    --scott

    --
    "C'est un Nagra. C'est suisse, et tres, tres precis."

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andy Burns@21:1/5 to Grant Taylor on Sun Jul 25 21:34:03 2021
    Grant Taylor wrote:

    I did my best to avoid PPPoE because of the complications related to MTU
    / MSS.

    I run openWRT on a VDSL capable router, specifically so I can force
    baby-jumbo on the VLAN of the WAN interface, to keep up at 1500 bytes ...

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Theo@21:1/5 to Scott Dorsey on Mon Jul 26 10:57:43 2021
    Scott Dorsey <kludge@panix.com> wrote:
    Grant Taylor <gtaylor@tnetconsulting.net> wrote:
    Where I'm from there is a lot of rural area where a single DSLAM may
    have trouble covering line length to connect to 32 houses. So 8 and 16 >port DSLAMs were common. In such sparse areas, each ISP deploying their >own DSLAM would be *EXTREMELY* inefficient if not wasteful.

    Where I am, we have low population density, but only one DSL provider,
    namely the telco that owns the local loops. So that's who owns the
    DSLAM and when you get out onto the vast and dizzying internet, you are
    doing so from an IP address owned by them. This is typical of much of
    rural America.

    In the UK, in more rural areas it's less common for non-BT ISPs to have
    DSLAMs at the exchange (central office). Exchanges are typically every few miles, and in particularly rural areas may serve few customers (probably >32 though, with a line length of a few km). So your line terminates in a BT Openreach-owned DSLAM, but then runs over a virtual tunnel to your ISP (I'm
    not sure if it's an ethernet VLAN or something else). Exchanges were
    typically sizeable windowless buildings with Strowger switches, nowadays
    they can be simple equipment cabinets by the roadside (particularly with
    VDSL where each cabinet becomes a mini exchange).

    You contract service from your ISP, who routes your packets onto the
    internet. Your ISP subcontracts your DSL local loop from BT Openreach. If
    you have a fault on your local loop you have to contact your ISP who raises
    a ticket with Openreach.

    Before VDSL, if your exchange became populated enough, your ISP may decide
    to install their own DSLAM, and so your loop is physically jumpered onto
    their rack rather than the Openreach one. That saves them some money as
    they don't have to rent Openreach's DSLAM (just their power and rack space)
    and they aren't running a million tunnels back to their datacentre as they
    can terminate directly at their equipment in the exchange. They then just
    have to rent fibre backhaul from your exchange to their datacentre.

    All of this is transparent to the customer - for anyone with an Openreach connection (almost everyone, since BT were the nationalised telco) they have
    a choice of ISPs. The details only matter to the speed you get, the price
    you pay, and some edge cases when switching from one ISP to another.

    Theo

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anssi Saari@21:1/5 to Spiros Bousbouras on Mon Jul 26 19:04:15 2021
    Spiros Bousbouras <spibou@gmail.com> writes:

    You tell it via the "default route" you setup when you configure it.

    I never did any such configuring on my current router. But , since it was provided by my ISP , it probably came preconfigured.

    There is a protocol commonly used for this sort of thing, it's called
    DHCP. Basically your router sends and receives broadcast packets and
    accepts any responses which contain an IP address and default
    route.

    DHCP's not the only option, PPP is also used apparently since others
    mentioned it. My country implemented ADSL usually with RFC1483 where
    ethernet packets are stuffed into ATM cells. Awful for efficiency, about
    15% wasted on framing total (TCP+IP+Ethernet+ATM headers). Although I
    assume PPP doesn't help with that much.

    But this leads to another question : what if someone enters a route
    not for an ISP they are subscribing to ?

    Your router only talks to the ISP's router and the ISP's router will
    likely drop such packets on the floor. Routers route what they're
    configured to route where they are configured to route it. And if no
    route exist, well, your bits aren't going to go anywhere.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Spiros Bousbouras on Mon Jul 26 13:47:02 2021
    On 7/25/21 11:42 AM, Spiros Bousbouras wrote:
    I never did any such configuring on my current router. But , since
    it was provided by my ISP , it probably came preconfigured.

    Yes, many ISPs that provide equipment tend to pre-configure as much of
    it as they can with reasonable effort.

    But this leads to another question : what if someone enters
    a route not for an ISP they are subscribing to ?

    I suspect that the worst that will happen is that things simply won't work.

    I'm guessing that some authentication will fail but if they somehow
    manage to crack the authentication process then they can get internet
    access without paying an ISP , right ?

    The nature of how ISP networks work tends to be myopic, despite
    technical capability for sharing.

    Chances are extremely good that, without someone at the ISP or telco reconfiguring something, there is nothing that an ISP subscriber can do
    to access another ISP's network. As in traffic will never make it from
    the circuits associated with ISP A over to ISP B.

    I was reading earlier en.wikipedia.org/wiki/TR-069

    I don't think that TR-069 will be of any use if there is not a
    functional communications path.

    and they say that routers are crackable.

    Almost always.

    The setting is different but if an outsider can crack a router then
    I'm guessing that the owner can too.

    Usually. Though I've found that often, external actors are more
    motivated than equipment owners.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Grant Taylor@21:1/5 to Anssi Saari on Mon Jul 26 13:58:21 2021
    On 7/26/21 10:04 AM, Anssi Saari wrote:
    DHCP's not the only option, PPP is also used apparently since others mentioned it.

    I think you mean PPP-over-Ethernet, or PPPoE.

    Yes, PPPoE uses PPP as the base protocol, but there is more to PPPoE
    that means that it can't be used interchangeable with (just) PPP.

    My country implemented ADSL usually with RFC1483 where ethernet
    packets are stuffed into ATM cells. Awful for efficiency, about 15%
    wasted on framing total (TCP+IP+Ethernet+ATM headers). Although I
    assume PPP doesn't help with that much.

    My understanding is that PPPoE also uses RFC 1483 Ethernet Bridging.
    It's TCP+IP+PPPoE+Ethernet+ATM in that case.

    There are other was to put IP, and likely PPP(oE) on ATM, but they are
    seldom done as (over) Ethernet has become the defacto standard.

    Your router only talks to the ISP's router ....

    Likely not even that.

    The OPs router will send frames into the ether looking for the router
    it's configured to talk to. It will extremely likely be unable to find
    said router, thus won't talk to any router.

    Routers route what they're configured to route where they are
    configured to route it. And if no route exist, well, your bits aren't
    going to go anywhere.

    Agreed. However, "default gateways" on the ISP's router tend to mean
    that they will route just about everything (unless they've been
    configured to reject something).

    Aside: "gateway" and "router" are the same thing in this context.
    Different terms that imply when the different terms for the same thing
    was used.



    --
    Grant. . . .
    unix || die

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anssi Saari@21:1/5 to Grant Taylor on Wed Jul 28 20:51:15 2021
    Grant Taylor <gtaylor@tnetconsulting.net> writes:

    There are other was to put IP, and likely PPP(oE) on ATM, but they are
    seldom done as (over) Ethernet has become the defacto standard.

    Your router only talks to the ISP's router ....

    Likely not even that.

    I was really trying to keep this mostly on the IP level since that's
    what the OP asked about. The physical level isn't really relevant to
    him. I guess I should've made that clear. And left out all the stuff
    about the low level stuff.

    The OPs router will send frames into the ether looking for the router
    it's configured to talk to. It will extremely likely be unable to
    find said router, thus won't talk to any router.

    I have no idea what you mean here. How does he get online then? And what
    does his router actually connect to then? On the IP level? In my
    understanding communication on IP level is computer to router to router
    to router to eventual server.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Bruce Horrocks@21:1/5 to Spiros Bousbouras on Wed Jul 28 21:37:25 2021
    On 25/07/2021 19:30, Spiros Bousbouras wrote:
    I'm asking for my general education so any reply for any location and any
    set up would be interesting to me.

    There's a good explanation of how BT implemented ADSL in the UK here. <https://kitz.co.uk/adsl/equip.htm>

    It's a bit dated in that it only covers up to ADSL but it's not wrong.

    The only area where it falls down a bit is that when it refers to ISPs
    it does so in terms of larger ISPs that provide their own kit. It's
    perfectly possible for a small ISP to buy in everything from BT or one
    of the larger ISPs (or both at the same time for different exchanges)
    and merely handle the billing relationship with the householder.

    Typically these are ones that want to concentrate on, for example,
    providing web hosting and business web services rather than being a
    'hardcore' ISP.

    Finally there is a last type of ISP, less common now, which is a purely
    virtual ISP run on behalf of a large supermarket chain or energy
    supplier / utility company.

    --
    Bruce Horrocks
    Surrey, England

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)