Printable View
IP datagrams contain a TTL (Time To Live) section that states how many hops the packet should be allowed to make before being killed. Does anyone know the maximum amazing of hops that can be stated ? .... and surely if it is high number couldn't a network be DoS'ed by letting off many packets with no destination and the maximum TTL number therefore creating massive amounts of traffic that just hops around the network ?
I know that if you use traceroute the maximum hops is usually around 30.
p.s. Is it possible to let of packets onto a network or the internet without a destination? Or do you mean a destination that is false, so that it keeps looking?
I think that maximum is 255... but can't be sure... default is 128
The TTL field in the IP Datagram header is one byte in size, giving it a maximum of 255 as was stated above. This field is the number of hops(routers) that a packet may traverse before it is dropped. This field was originally the number of seconds that a packet would stay alive before it was dropped, however, since it was rare for a packet to remain at the router for more than 1 second, and the routers did not have access to a common clock, it became standard practice to simply decrement the value by one.
The default value of TTL was 64 in 2000, this may have changed
A RARP request, frame, has no destination IP address when sent. It contains only the MAC address of the source machine and is used to obtain the corresponding IP address from the RARP servers ARP table.Quote:
Originally posted here by Modderfokker
p.s. Is it possible to let of packets onto a network or the internet without a destination? Or do you mean a destination that is false, so that it keeps looking?
It is not possible to send an IP packet as two of the mandatory fields of a packet are the source and destination IP addresses, either a valid IP address, or a broadcast address.
One interesting thing is to measure how "wide" the internet is; the maximum numbers of hops needed to reach the two points most apart. Of course this cannot be done exactly but I think it was microsoft who did som testing once. And the result is that it's more than 32 but less than 60 (Probably why the TTL was set to 64 in w2000). This I was told in a lecture in Internet Technology, so unfortunately I can't give any references to any fact pages. The lecturer (who has been in the buisniess forever more or less) made a qualified guess of around 40 as the most needed hops.
So the maximum value allowed in a IPV4 datagram; 255 will be sufficent for a long long time.
The Hop Count field in IPV6, which is the corresponding field to the TTL-field in IPV4, is also one byte large, so the designers also probably made the assumption that 255 will be sufficent.
Skarlman > Something along those lines was defined in RFC1122:
Quote:
A fixed value must be at least big enough for the Internet "diameter," i.e., the longest possible path. A reasonable value is about twice the diameter, to allow for continued Internet growth.