|
|
For those wishing to delve deeper into the possible range of DSL services, here's an explanation of some of the possibilities
By John Kemp (kemp@ns.uoregon.edu)
Another differentiating factor between ISPs is whether they accept DSL traffic that is in "bridged" mode or in "PPP" mode. The primary differences between these two modes of operation are explained below.
Bridged Mode. In bridged mode, the computer is assigned an IP address, but the packets from the customer's modem are dumped directly on the ISP's Local Area Network (LAN). The advantage of this technique is that it makes the customer appear to be local; that is, services that require broadcasting or non-IP protocols can operate as they normally would, assuming the ISP accepts or routes any of these protocols. The disadvantage is that the ISP then has to process all of this extra traffic, so this can place a greater load on the ISP's LAN.
PPP Mode. Things are not quite so straightforward in PPP mode. In this mode, the DSL modem is assigned an IP address and a route to an IP address at the ISP. Typically, the customer's computer also requires an IP address. In this case, the data that's sent to the ISP must first be routed by the DSL modem. Most ISPs will only accept IP traffic in this configuration; neither broadcasts nor other non-IP protocols will be allowed. The advantage is that the ISP is not burdened with broadcast traffic or the handling of non-IP protocol traffic, so their overall LAN performance may improve. The disadvantage, generally, is the extra complexity involved in the IP address configuration of the customer's computer and DSL modem.
In the general case, the choice between "bridged-mode" and "PPP-mode" would be of little concern to the average DSL user. But if you choose USWest as your ISP, in the service offerings that are of lower cost you will not only be required to use PPP, you'll also be required to use something called "NAT." NAT stands for Network Address Translation. It's a method of conserving IP addresses by using nonroutable "private" IP addresses on internal computers. NAT can also be used as a method for "hiding" internal computers from remote connections to make them more secure.
Drawbacks. As you might expect, the NAT method sacrifices some usability. The DSL modem can easily translate requests originating from the customer site going to outside services. When a connection is initiated from the customer site, the DSL modem automatically translates from the private address of the customer computer to the DSL modem address--the public address that the rest of the world sees.
But for connections initiated from a remote site, the private address assigned to the customer computer is nonroutable by anything other than the customer's own DSL modem. So it becomes difficult to offer services from the customer site to remote users. Since the customer's computer address is nonroutable, the remote site has no way to make the initial connection to the customer address (for example, hosting a web server or setting up an FTP site becomes a real challenge).
NAT can also impact many of the more complex IP network protocols that carry IP addresses in them as part of their data. NetMeeting conferencing protocol is one such example. PPTP, a security protocol, is another example of a protocol that has trouble working with NAT. NAT also performs poorly on services that are UDP based (i.e., not connection oriented), such as IRC servers and game servers.
So while NAT can be good for overall network performance and good for those who want the added security of a nonroutable IP address, it can impose a burden on those who want to offer services from their computer. Users who want to offer web pages or an FTP server would be challenged. And the loss of the ability to use services like NetMeeting, IRC, or game servers can be significant to some users.
A number of workarounds or alternatives to NAT do exist. The most obvious of these is to choose one of the higher-cost ISP service packages from USWest that provides static routable IP addresses. USWest classifies these service offerings, which do not involve the use of NAT at the customer site, as "Office" or "Business" packages.
Another alternative is to choose an ISP that offers bridged-mode service. Some ISPs offer both bridged-mode service and a static routable IP address for the customer computer as part of their base offering. This can be an attractive option, assuming the ISP also provides adequate service in other areas, such as routing topology and overall bandwidth. The more savvy user can make modifications to the settings of the DSL modem to allow NAT to successfully pass specific protocols. People refer to this technique as "setting NAT rules." Because this process can sometimes be complex, and because even the IP address of the modem may change over time, this is not the easiest approach and may not always achieve the desired goals.
Another way to get around NAT is to purchase the Cisco 605 internal DSL modem card instead of the Cisco 675 external modem. The use of the internal modem card eliminates the need for NAT, since the card itself handles all of the network traffic. While the IP address used by the card may still change from time to time, NAT will never be involved. The use of an internal card does involve some trade-offs, however. You may not want to attempt this method at a site where more than one computer will be connected.