TCP/IP: A great highway but dangerous to boot

Opinion
Jan 24, 20082 mins

* TCP/IP: Connection-oriented vs. connectionless

In the last newsletter, we began a discussion of TCP/IP and its 25th anniversary. One of the major shifts that IP brought about, of course, was the move from connection-oriented to connectionless protocols.

Of course, the fundamental difference between connection-oriented and connectionless is that a connection-oriented protocol first established a path and then the subsequent packets follow the same path (much like a phone call). By contrast, a connectionless protocol avoids the call setup process, but by doing so there is a need for a much larger address space on each packet.

So why does the address space matter?

It doesn’t now, but it sure did 25 years ago. Connectionless protocols have massive overhead on a per-packet basis as compared with connection-oriented protocols. Not only must the destination address on every packet be unique throughout the entire network, but the sender’s address must also be included. Until relatively recently, it was not a safe assumption that WAN bandwidth is sufficiently inexpensive to make the overhead a “don’t care” issue. In fact, this is an assumption that could only be made within roughly the past five years.

We’ve always said that one of the best things about IP is that it is a wonderfully simple and open protocol. We’ve also always said that one of the worst things about IP is that it is a wonderfully simple and open protocol. This has resulted in the sad state of affairs that if there was a slightly flawed assumption in the development of TCP/IP, it was an assumption that the users would follow the rules. And this was not a bad assumption when TCP/IP was for internal use within DARPA.

However, once popularized and opened to the community in general, the many ways to abuse IP (and related protocols like SMPT) have been and continue to be exploited. As a result, the introduction of next generation firewalls (as discussed last week) has become a necessity.

In summary, the great thing about TCP/IP is that it’s a great highway. But there are a heckuva lot of dangerous drivers on it.

Jim has a broad background in the IT industry. This includes serving as a software engineer, an engineering manager for high-speed data services for a major network service provider, a product manager for network hardware, a network manager at two Fortune 500 companies, and the principal of a consulting organization. In addition, Jim has created software tools for designing customer networks for a major network service provider and directed and performed market research at a major industry analyst firm. Jim’s current interests include both cloud networking and application and service delivery. Jim has a Ph.D. in Mathematics from Boston University.

More from this author