It is hard to think of any real drawbacks. Generally it is not configured to allocate the same IP address to the same device, but this can be done, so it can even be used for devices such as network printers. Although it is usually considered good practice to configure the address on the device itself rather than use DHCP. I suppose that DHCP servers could be faked by someone wanting to disrupt the network, but that is not a hugely likely event. A possible drawback is that it requires some knowledge of how to configure it to work correctly, and how to configure routers to pass the DHCP requests to the server when the client is on a different subnet.
Overall I would say that the benefits vastly outweigh the drawbacks (which are few).
But I am sure there is some good mileage in this discussion
DHCP is a broadcast protocol. When a client asks for IP address info, all systems have to process the packet. So if the lease duration is set to a very short time, the network could become flooded by DHCP request broadcasts. The challenge with DHCP is address to name resolution. This means that DHCP cannot stand alone without dynamic DNS. Not all DHCP is created the same.
Yes, rogue DHCP could be a real issue if someone wanted to publish a malicious DNS server for client name resolution. The DNS server could resolve microsoft.com to malwaresoftware.com or something. A rogue DHCP server could even pass out bad network information and cause clients to not be on the same subnet as company resources. I have seen this happen if someone brings in a rogue wireless access point with DHCP on the LAN side handing out a 192.168.x.x subnet when the clients should be on a company subnet. This caused the clients who got addresses from this rogue DHCP server to “lose connectivity” with the resources they needed to access. This could be considered a denial of service (DOS) attack.
In the IT trenches? So am I – read my IT-Trenches blog