- What exactly do you want to accomplish by linking your network with some other organization's network? The answer to this question will determine what services you need to provide (and, by implication, what services should be blocked).
- Are you just looking to exchange email or files with the other organization privately, without having to communicate over the Internet? If that's all you want, then maybe a dial-up UUCP connection is all you need, not an IP-level connection between your nets.
- Are you trying to create a full work environment for a joint project in which team members from both organizations can work together and yet still have access to their own "home" systems (which need to be protected from the other organization)? In such a case, you might actually need two firewalls: one between the joint project net and each of the home organizations.
- Are you looking for something in between? Exactly what you're trying to accomplish, and what your security concerns are, will determine what firewall technologies are going to be useful to you.
An 'arms-length relationship': shared-perimeter networks
Shared perimeter networks are a good way to approach joint networks. Each party can install its own router, under its own control, onto a perimeter net between the two organizations. In some configurations, these two routers might be the only machines on the perimeter net, with no bastion host. If this is the case, then the "net" might simply be a high-speed serial line (such as a 56-kilobit-per-second line or T1/E1 line) between the two routers, rather than an Ethernet or another type of local area network.
This is highly desirable with an outside vendor. Most of them are not networking wizards, and they may attempt to economize by connecting multiple clients to the same perimeter network. If the perimeter net is an Ethernet or something similar, any client that can get to its router on that perimeter network can see the traffic for all the clients on that perimeter network--which, with some providers, is almost guaranteed to be confidential information belonging to a competitor. Using a point-to-point connection as the "perimeter net" between the outside vendor and each client, rather than a shared multiclient perimeter net, will prevent them from doing this, even accidentally.
Do you need bastion hosts?
You might not actually need to place a bastion host on the perimeter network between two organizations. The decision about whether you need a bastion host depends on what services are required for your firewall and how much each organization trusts the other. Bastion hosts on the perimeter net are rarely required for relationships with outside vendors; usually you are sending data over one particular protocol and can adequately protect that as a screened host.
If the organizations have a reasonable amount of trust in each other (and, by extension, in each other's security), it may be reasonable to establish the packet filters so that clients on the other side can connect to internal servers (such as SMTP and DNS servers) directly.
On the other hand, if the organizations distrust each other, they might each want to place their own bastion host, under their own control and management, on the perimeter net. Traffic would flow from one party's internal systems, to their bastion host, to the other party's bastion host, and finally to the other party's internal systems.
What the future holds
Systems that might be called "third-generation firewalls" -- firewalls that combine the features and capabilities of packet filtering and proxy systems into something more than both -- are just starting to become available.
More and more client and server applications are coming with native support for proxied environments. For example, many WWW clients include proxy capabilities, and lots of systems are coming with run-time or compile-time support for generic proxy systems such as the SOCKS package.
Packet filtering systems continue to grow more flexible and gain new capabilities, such as dynamic packet filtering. With dynamic packet filtering, such as that provided by the CheckPoint Firewall-1 product, the Morning Star Secure Connect router, and the KarlBridge/KarlBrouter, the packet filtering rules are modified "on the fly" by the router in response to certain triggers. For example, an outgoing UDP packet might cause the creation of a temporary rule to allow a corresponding, answering UDP packet back in.
The first systems that might be called "third generation" are just starting to appear on the market. For example, the Borderware product from Border Network Technologies and the Gauntlet 3.0 product from Trusted Information Systems look like proxy systems from the external side (all requests appear to come from a single host), but look like packet filtering systems from the inside (internal hosts and users think they're talking directly to the external systems). They accomplish this magic through a generous amount of internal bookkeeping on currently active connections and through wholesale packet rewriting to preserve the relevant illusions to both sides. The KarlBridge/KarlBrouter product extends packet filtering in other directions, providing extensions for authentication and filtering at the application level. (This is much more precise than the filtering possible with traditional packet filtering routers.)
While firewall technologies are changing, so are the underlying technologies of the Internet, and these changes will require corresponding changes in firewalls.
The underlying protocol of the Internet, IP, is currently undergoing major revisions, partly to address the limitations imposed by the use of four-byte host addresses in the current version of the protocol (which is version 4; the existing IP is sometimes called IPv4), and the blocks in which they're given out. Basically, the Internet has been so successful and become so popular that four bytes simply isn't a big enough number to assign a unique address to every host that will join the Internet over the next few years, particularly because addresses must be given out to organizations in relatively large blocks.
Attempts to solve the address size limitations by giving out smaller blocks of addresses (so that a greater percentage of them are actually used) raise problems with routing protocols. Stop-gap solutions to both problems are being applied but won't last forever. Estimates for when the Internet will run out of new addresses to assign vary, but the consensus is that either address space or routing table space (if not both) will be exhausted sometime within a few years after the turn of the century.
While they're working "under the hood" to solve the address size limitations, the people designing the new IP protocol (which is often referred to as "IPng" for "IP next generation" -- officially, it will be IP version 6, or IPv6, when the standards are formally adopted and ratified) are taking advantage of the opportunity to make other improvements in the protocol. Some of these improvements have the potential to cause profound changes in how firewalls are constructed and operated; however, it's far too soon to say exactly what the impact will be. It will probably be at least 1997, if not later, before IPng becomes a significant factor for any but the most "bleeding edge" organizations on the Internet. (Chapter 6 of our book describes IPv6 in somewhat more detail.)
The underlying network technologies are also changing. Currently, most networks involving more than two machines (i.e., almost anything other than dial-up or leased lines) are susceptible to snooping; any node on the network can see at least some traffic that it's not supposed to be a party to. Newer network technologies, such as frame relay and Asynchronous Transfer Mode (ATM), pass packets directly from source to destination, without exposing them to snooping by other nodes in the network.