About IP Addresses - TCP-IP

This is a discussion on About IP Addresses - TCP-IP ; Hi, I'm developing an online game that logs the names each player uses (using their IP address as the key). The idea behind this is that if a known person were to start using an unknown name, it would be ...

+ Reply to Thread
Results 1 to 9 of 9

Thread: About IP Addresses

  1. About IP Addresses

    Hi,

    I'm developing an online game that logs the names each player uses (using
    their IP address as the key). The idea behind this is that if a known
    person were to start using an unknown name, it would be possible to
    determine who they are by viewing their 'name history'.

    This works as follows. The game assumes that if Player X is using an IP
    address that has been used before on the server (by Player Y), then Player
    X is using the same ISP account as Player Y. It also assumes that if
    Player Z is using an IP address that has never been used on the server,
    then Player Z is using an ISP account that has never been used on the
    server.

    This isn't infallible, of course, but it works quite well in practise.

    Now, from what I've seen, the following actually holds true: If any two IP
    addresses are identical in their first three octects (a.b.c.*) then, even
    if they differ in their last octet (d), they represent the same ISP
    account.

    My question is: Is this a fair assumption? What, roughly, is the
    likelihood that it's true?

  2. Re: About IP Addresses

    In article ,
    tjb wrote:

    >Now, from what I've seen, the following actually holds true: If any two IP
    >addresses are identical in their first three octects (a.b.c.*) then, even
    >if they differ in their last octet (d), they represent the same ISP
    >account.


    >My question is: Is this a fair assumption?


    I'm not sure what you mean by 'ISP account'. If you mean 'ISP'
    that's one thing, but if you mean the same user (e.g., the same residence)
    of a particular ISP then No, your assumption is completely wrong.

    In the case of business accounts, the last octet might be officially
    divided between 1 to 64 -different- businesses. (Look up "SWIP".)

  3. Re: About IP Addresses

    Walter Roberson wrote:
    >>Now, from what I've seen, the following actually holds true: If any two IP
    >>addresses are identical in their first three octects (a.b.c.*) then, even
    >>if they differ in their last octet (d), they represent the same ISP
    >>account.

    >
    >>My question is: Is this a fair assumption?

    >
    > I'm not sure what you mean by 'ISP account'. If you mean 'ISP'
    > that's one thing, but if you mean the same user (e.g., the same residence)
    > of a particular ISP then No, your assumption is completely wrong.


    Yes -- the same user.

    > In the case of business accounts, the last octet might be officially
    > divided between 1 to 64 -different- businesses. (Look up "SWIP".)


    Would that matter for games though? I'd guess not, but then I don't know
    much about it. If one were to factor that out, would the assumption in
    question still be a bad one?

    Thanks!

  4. Re: About IP Addresses

    In article <1kim7gd96v2xt.dlg@tjb.invalid.invalid>,
    tjb wrote:
    >Walter Roberson wrote:
    >>tjb wrote
    >>>Now, from what I've seen, the following actually holds true: If any two IP
    >>>addresses are identical in their first three octects (a.b.c.*) then, even
    >>>if they differ in their last octet (d), they represent the same ISP
    >>>account.


    >>>My question is: Is this a fair assumption?


    >> I'm not sure what you mean by 'ISP account'. If you mean 'ISP'
    >> that's one thing, but if you mean the same user (e.g., the same residence)
    >> of a particular ISP then No, your assumption is completely wrong.

    >
    >Yes -- the same user.


    In that case, NO, your assumption is completely unsafe.

    Residential DSL accounts change IPs often, sometimes once a week,
    and re-use of the final octet is part of the technology. If your server
    is popular enough to worry about multiple use of the same accounts
    then the likelyhood of an IP clash approaches unity over time.

    If there were an ISP that dynamically allocates user addresses
    from a /16 (65536 addresses), then if you had just 151 users from
    that ISP and they were to change their address -once- (e.g., at the
    end of the week when the ISP forced the IP change), then the
    probability would be over 50% that one of the IPs would get reused.
    Divide that number by the number of weeks of IP history you intend
    to keep (though there would be a small correction due to the possibility
    that the duplication was from a particular IP/user combination to
    the same IP/user combination.)

  5. Re: About IP Addresses

    In article <1kim7gd96v2xt.dlg@tjb.invalid.invalid>,
    tjb wrote:

    > Walter Roberson wrote:
    > >>Now, from what I've seen, the following actually holds true: If any two IP
    > >>addresses are identical in their first three octects (a.b.c.*) then, even
    > >>if they differ in their last octet (d), they represent the same ISP
    > >>account.

    > >
    > >>My question is: Is this a fair assumption?

    > >
    > > I'm not sure what you mean by 'ISP account'. If you mean 'ISP'
    > > that's one thing, but if you mean the same user (e.g., the same residence)
    > > of a particular ISP then No, your assumption is completely wrong.

    >
    > Yes -- the same user.


    Then you're totally wrong. For instance, most cable modem ISPs assign a
    /22, /23, or /24 block (i.e. 256, 512, or 1024 contiguous addresses) to
    a cable node. So addresses with the same first three octets, but
    different last octets, will usually be customers in the same
    neighborhood, but NOT the same customer.

    It sounds like you're assuming that you won't have many players in the
    same neighborhood, so if you see a.b.c.x and later a.b.c.y, that it's
    more likely that it's the same user who got assigned a new IP in the
    same block, rather than one of his neighbors.

    --
    Barry Margolin, barmar@alum.mit.edu
    Arlington, MA
    *** PLEASE post questions in newsgroups, not directly to me ***
    *** PLEASE don't copy me on replies, I'll read them in the group ***

  6. Re: About IP Addresses

    "tjb" pondered:
    > I'm developing an online game that logs the names each player uses
    > (using their IP address as the key). The idea behind this is that if a
    > known person were to start using an unknown name, it would be
    > possible to determine who they are by viewing their 'name history'.


    What if two customers on two different computers both have the
    same IP address? That is, what if they are connected through the
    same Internet Service Provider and running on the same ISP
    account? This was possible when everyone used phone modems
    and with high speed connections tends to become common, esp.
    for businesses and geeks.

    Effectively, they could be running behind a proxy to get such a
    configuration. Microsoft has provided the Proxy abilities free of
    charge for quite a long time. You should be looking at MAC
    addresses as well. And you may even find that one person with
    two Gaming accounts might log in with the SAME IP address
    and either two different MAC addresses or perhaps the same
    MAC address. Two MAC addresses indicates they have two
    network cards on the system. There might be other ways to
    accomplish such things, but these represent the ways I've done
    such things in the past. In the case if the same MAC address
    and the same IP address, for the server side, I'd create a
    unique ID (autoincrementing for each connection that comes
    in). And then store the gaming UserName as a string field for
    each connection (this assumes the UserNames are unique and
    represents the way to identify each individual in your gaming
    world).

    Good luck.

    --
    Jim Carlock
    Post replies to the group.



  7. Re: About IP Addresses

    In article <3b3Qg.188$Rr.149@tornado.tampabay.rr.com>,
    "Jim Carlock" wrote:

    > "tjb" pondered:
    > > I'm developing an online game that logs the names each player uses
    > > (using their IP address as the key). The idea behind this is that if a
    > > known person were to start using an unknown name, it would be
    > > possible to determine who they are by viewing their 'name history'.

    >
    > What if two customers on two different computers both have the
    > same IP address? That is, what if they are connected through the
    > same Internet Service Provider and running on the same ISP
    > account? This was possible when everyone used phone modems
    > and with high speed connections tends to become common, esp.
    > for businesses and geeks.


    I think this issue was alluded to in his original post, where he wrote:

    > This isn't infallible, of course, but it works quite well in practise.


    He's simply assuming that there won't be two users behind the same proxy
    or NAT router.

    --
    Barry Margolin, barmar@alum.mit.edu
    Arlington, MA
    *** PLEASE post questions in newsgroups, not directly to me ***
    *** PLEASE don't copy me on replies, I'll read them in the group ***

  8. Re: About IP Addresses

    Hello,

    Jim Carlock a écrit :
    >
    > addresses as well. And you may even find that one person with
    > two Gaming accounts might log in with the SAME IP address
    > and either two different MAC addresses or perhaps the same
    > MAC address.


    You can't see MAC addresses across the internet.

  9. Re: About IP Addresses

    In article ,
    Pascal Hambourg wrote:

    >> addresses as well. And you may even find that one person with
    >> two Gaming accounts might log in with the SAME IP address
    >> and either two different MAC addresses or perhaps the same
    >> MAC address.

    >
    >You can't see MAC addresses across the internet.


    You don't see MAC addresses accross the Internet in link layer headers,
    but you can see them or values that depend on them in IP payloads.
    A now classic example is the mechanism that allowed the arrest and
    conviction of some Microsoft virus or worm authors. I think another
    example is the mechanism that Microsoft uses to authorize installations
    of Windows XP.


    Vernon Schryver vjs@rhyolite.com

+ Reply to Thread