> >Hello all,
> >I am using Windows NT as a DHCP server with a range of IP addresses to
> >issue say 184.108.40.206 to 220.127.116.11. I have a Win95 client with
> >hardcoded IP address of 18.104.22.168. Also, I have another Win95
> >client which is configured to obtain an IP address from the network
> >using DHCP.
> >I booted up the Windows NT server and the Win95 client with IP address
> >22.214.171.124 first. Then, I boot up the Win95 DHCP client and got an IP
> >address of 126.96.36.199 issue from the Windows NT DHCP server. I got
> >complaint message of duplicate IP address from both Win95s.
> >Is this a bug in Windows NT as the algorithm is not according to the
> >RFC 1531(DHCP). The DHCP server is supposed to send an ICMP echo to
> >whether anyone using that IP address before issuing out rite?
> >Any comments or feedback?
> The RFC doesn't require the server to figure out whether there are
> other machines using addresses within its scope...statically assigned
> addresses musn't overlap the dynamically assigned range.
Yes, we understand the RFC and the issue is not only how to detect the
duplicate ip address but, shouldn't the dynamic allocation of network
addresses mechanism of DHCP server allocate those unassigned ip addresses
within the range and yet not used in that subnet to the newly join
> MS TCP stacks do perform duplicate address detection when they
> initialize, regardless of whether they were dhcp configured, or
> statically configured. That's why the client complains.
How about some proactive measure (for large network) with "highly
computer literate" end-users? I think M$ should produce a more "foolproof"
> Dave MacDonald
> Microsoft Corporate Network Systems
> The opinions expressed in this message are my own personal views
> and do not reflect the official views of Microsoft Corporation.