Date: Sat, 17 Nov 2001 00:10:14 -0800 From: "Ted Mittelstaedt" <tedm@toybox.placo.com> To: "Anthony Atkielski" <anthony@atkielski.com>, "Bara Zani" <bara_zani@yahoo.com>, <freebsd-questions@FreeBSD.ORG> Subject: RE: DSL PPPoE with 2 NICs Message-ID: <000a01c16f3f$497965c0$1401a8c0@tedm.placo.com> In-Reply-To: <000a01c16e97$fa4e3130$0a00000a@atkielski.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>-----Original Message----- >From: Anthony Atkielski [mailto:anthony@atkielski.com] >Sent: Friday, November 16, 2001 4:12 AM >To: Ted Mittelstaedt; Bara Zani; freebsd-questions@FreeBSD.ORG >Subject: Re: DSL PPPoE with 2 NICs > > > >> But this is because businesses have to make tradeoffs >> all the time, and a very common tradeoff is giving >> up functionality for a cheaper cost. > >Giving up functionality you don't need is irrelevant; and if you need it, you >cannot give it up. The business requirement is to solve the problem, not to >please the geeks. > What the best long term solution is and what is needed to solve the immediate business problem are often two different things. I'll cite an example from the routing game. (names are of course bogus to protect the indecent) A small manufacturing firm named "West Electronics" finally decides to connect a dedicated line to the Internet. They decide that they are going to start out with a DSL circuit. They purchase a Linksys and install it. 6 months later they decide that they want to set up a webserver and the DSL line isn't going to cut it, so they purchase a point-to-point T1. Since the Linksys can't handle that they chuck it and buy a Netopia router. One year later they find they need redundancy and so they chuck the Netopia and buy a Cisco 1604 and a T1 card for it. 6 months after that they decide they need a firewall so they go out and buy a Netscreen and put it between the 1604 and the inside net. At the same time, another manufacturing firm named "East Electronics" decides to connect a dedicated line to the Internet. They decide that they are going to start out with a DSL circuit. They purchase a Cisco 2620 dual-port Ethernet router and install it. 6 months later they decide that they want to set up a webserver and the DSL line isn't going to cut it, so they purchase a T1 card and slip it into the 2620. One year later they find they need redundancy and so they purchase an ISDN BRI card and slip it into another slot on the 2620. 6 months after that they decide they need a firewall so they purchase the IOS/Firewall feature set and load it on the 2620. Now, East Electronics spent more money initially on the 2620. But, overall, they are much more secure now than West Electronics. This is because the admin at East Electronics has had 2 years of experience with the Cisco 2620 and now knows it very well and has even taken some Cisco courses. By contrast, while West Electronics spent less initially, they are worse off because their admin has never had any vested interest in learning anything about their routing gear because it changes all the time, and so doesen't know how to properly set up an access list on the Netscreen because he doesen't understand access lists. You will find as you continue on your career, Anthony, that if you always base your purchasing decisions on the cheapest solution at the time, that you will end up paying for it because your going to be forever tearing out your existing inadequate systems and replacing them. Thus, all institutional knowledge that you gain is wasted. The majority of business in the US operate like this and when you get into the larger businesses you would be astounded at the amount of money that gets thrown down the rathole in failed IT projects. If, however, you look at some of the technological leaders you will find that they don't operate this way, instead they will invest more money in a more expensive solution upfront because they know that over time the solution is expandable. >> If the functionality that is given up is not essential >> to the operation of the business then most businesses >> have a fudiciary duty to their stockholders to go with >> the cheaper solution that has less features. > >In other words, the cheaper idea is the better idea, in contrast to what you >initially asserted. > Having a fudiciary duty and actually executing on it are two different things. For instance, all CEO's have a fudiciary duty to keep the corporation on solid financial footing and not sacrifice the corporation's profits for their own personal gain. However, in actuality you will find that very, very few CEO's of larger businesses actually act this way, most of them secure their own financial futures first and the company's second. In general, businesses do have to consider the cheaper solution as the "better" solution. If you never do this, your going to run out of money and go bankrupt. But, on the other hand, if you ALWAYS do this, your going to make poor decisions a lot of times. Sure, these are conflicting goals. The winners in business are those people that have developed a sense of when it's time to break the rules and invest the time and money upfront in the more expensive solution that's more expandable. This is what risk is all about. >> However, this kind of short term thinking sometimes >> will turn around and bite the business on the butt >> and cost it more in the long run, because they >> have to scrap the solution they bought 6 months >> ago and replace it with a more expensive solution >> that has the features that they suddenly found out >> they needed to have. > >No. The cost of a $100 router will be amortized in far less than six months. >It is cheaper to buy the router while it meets requirements, then >remove it if >necessary and replace it with a more elaborate solution later, But the problem here as others have pointed out is that you can only know with certainty if this is the right decision in HINDSIGHT. If that $100 router has a hole in it and someone guns you, then this is when your going to learn that it's necessary to remove it and replace it with a more elaborate solution later. Of course, it may so happen that the $100 router doesen't have a hole in it and nobody guns you and the future proceeds according to your plans. But, if it DOES then your going to spend a lot more money on putting everything back together than if you had gotten the more elaborate solution in the first place. > >Most businesses don't have emotional attachments to their network >architectures, >and they'll prefer my statement just because it gets to the point >and solves the >problem, rather than wax philosophical in an attempt to conceal a religious >preference for a specific configuration. > Both statements get to the point and solve the problem. But yours is easier to understand for the feeble-minded than mine. That's the only difference. >> You need to explore other editors under FreeBSD, >> most of them are much easier to use than vi. > >What do you recommend? I want something that works just like Notepad. > Well obviously if your Telneted in to a command prompt you can't run an editor that looks like Notepad because there is no mouse at your disposal. But, you can run the "ee" editor just by typing "ee" and have a much easier to use editor. If all you need to do is change a single byte in a config file it's a lot easier to use this while at a shell prompt than to FTP the file over then back. > >They are concerned because UNIX is not a router, and makes a poor >substitute for >one where there is a choice. It's rather like writing a C++ program >to change >the contents of a specific text file, instead of just calling the >file up in an >editor and changing it directly. > It would then probably come to you as a surprise that when TCP/IP was invented that the only kind of routers there were were UNIX systems. Specialized hardware routers came about later, and to this day there's still many, many ISP's running BGP4 on PC's with serial lines coming in to them. Hardware routers solve some unique things that PC's can't, like survival in harsh environments, for example the Cisco 1000 and 1600 series have no moving parts and can withstand higher temperatures and dust and vibration that would wreck a PC in short order. They also can be made to run off -48v DC which is often the only thing available in Telco closets. But if the environment is some closely-monitored room-temperature office environment, why then a PC is not a particularly poor substitute. You could do a lot worse, for example an Osicom hardware router. (which is an orphaned product) Ted Mittelstaedt tedm@toybox.placo.com Author of: The FreeBSD Corporate Networker's Guide Book website: http://www.freebsd-corp-net-guide.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000a01c16f3f$497965c0$1401a8c0>