A cheap gateway/firewall could provide security for you and client
Disclaimer: I am new to Linux but not to computers and computer science. Although I have an engineering degree and 40 years experience with computers of all shapes and sizes, at times it seems like no more than a year's experiance--40 times! So much too learn, so little time! So I have a few more opinions (prejudices) I carry as an extra burden than most.
I have followed interesting discussions on setting up cheap or free (386,486) Linux boxes on a private (192.168.x or other) net. Experimenting on my own time, I had much more trouble configuring older hardware than I anticipated.
Armed with this experience and, having a few years experience working for an ISP as a unix admin supporting thousands of customers running Macs and Windows I want to provide this service. But...There are all areas where my judgement tells me, be careful, you have a reputation to make, or lose...
A few things hold me back.
1. Training burden. Having become complacent with years of 'easy to use and consistent user interface' Mac and PC software I consistently find things more complex to set up and explain to a person new to Linux than I expect. These years of easy configuration have contributed to what I call a 'command line disability'. It is really an attitude that things maybe should be simpler than they are. You can't run a flexible, configurable and simple system unless you learn at least one way to do it from top to bottom without having to ponder each move. Desktop systems have made a market by making it possible to do most things without memorizing them and at least being able to search through similar menu items when you forget. Linux Desktops do this too, but not as consistently. It has been said that if you know how to use a Mac, you know how to use 80% of any Mac program. If you know how to use info you can find out about any Linux program too but I get distracted because of so many options. So I expect others with less experience and less commitment may have trouble finding a solution to a given problem. They might be intimidated and give up because it is too complex. I need a way to protect and educate these people while they are getting used to Linux or I will not be worth the money to them.
2. Expertise. To me it is essential to be comfortable with the command line if I expect to provide useful customer service. To me it is unreasonable to expect a customer to be comfortable with it. Configuring a Linux system for me at my present level of expertise is time consuming, humbling, worthwhile yet dollarwise cheap learning experience. However, the customer will not want to pay me for the learning experience nor suffer the lost time in exploring all options. One of the most appealing features of Linux to me is the idea that giving away Linux information is the best way to learn it. Once I learn it, it seems that I should be able to sell it in good conscience because the information will be reliable and applicable to my customer's needs. I think the market still demands a lot of information to set up and maintain secure, reliable internet service. The fact that the information is freely available and that the hardware is cheap doesn't mean it is easy. It just makes it feasible...for some of you.
3. Choice. You all know that in comparison to desktop machines which generally have few ways or one to do it, in *nix there is more than one way to do it. And since there are so many rainbows to chase, unless I severely limit my options, I end being different without being usefully different. I am learning to be careful with the options I select. I expect that the potential customers I will have will have a lesser tolerance than I for complexity. To provide a valuable service, it will be necessary to know the options I provide backwards and forwards.
On the positive side, several things propel me forward.
1. Really Cheap hardware. The cost of obtaining a Linux box capable of supporting a small nework should be a small percentage of the cost to the customer. I recently got an excellent Dell Pentium 100 off ebay for less than the cost of shipping it! I expect that it will be reasonable to be able to give credit for 'qualified trade-ins' that customers currently own that can provide reliable service with little more than a network card and modem upgrade. I just have to make sure I know how to make the hardware work with Linux and that it is solid. This should be great customer flexibility and value.
2. Marketing and Control. I can locate my hardware at the customer site and provide the connection service for a monthly fee. I recently worked for an ISP that provided a dialup router that cost $800 free if they signed for 3 months of $350/month ISDN24x7 dual channel service. This took a lot of the risk out for the customer because the ISP provided the ISDN connection at the customer site. The customer was not able to make any changes to the router, not even a password. The customer connection was continuously monitored 24x7 to make sure it passed traffic. This protected the ISP from unnecessary maintenance problems. And there were none. In comparison to dialup customers, these customers were almost maintenance free. For this business class of microbusiness customers there were different expectations from ordinary dialup. They ran their own small networks and did not want a lot of options.
3. Secure Remote Administration. Telnet (ssh) provides a secure, practical, flexible way to maintain a network via the internet or, for more security conscious folks, dial-in. Since I have control over the gateway/firewall at the customer site, I can provide good security at minimal cost.
4. Internet backups. The cost of internet storage should make it very practical to back up the most critcal files for small networks.
5. Very inexpensive support. The very nature of open source is such that the risk of not being able to solve a problem is vanishingly small if you stay with what you know. But I do not always have to be the person who solves the problem. Many online services provide me with support on subjects where I might have screwed up something I missed. All I really need to do is to be able to clearly state the problem. Then it is simply a matter of time before someone is able to point me to a solution at very little expense. This will end up saving my customer when I fall short. (Will those who have never fallen short please step this way? We have a cross for you to bear!) My experiences getting support with commercial providers has been expensive and generally not as capable. This is excellent assurance for my customer that I will not fail him. I am literally supported by a cast of thousands!
6. Minimal travel expense. I figure it costs me about $1/per mile distance I have to travel to a customer site. If I can set it up so that maintenance conversations are either through a local call or via the internet, I am saving the customer and myself money, time and inconvenience.
7. More timely and effective support during business hours. My experience supporting customers for an ISP tells me that remote support is superior in many ways to onsite support as long as both the customer and the support person can see what is happening. The ability to work cooperatively with a business customer is a tremendous business asset for both me and the customer. It is ideal for training.
8. Opportunity for value-adds to their network. Adding network resources can result in savings that you can share with the customer. More capability for the same money.
9. Internet in a box. A few years ago, CDs were marketed as Internet in a Box. They allowed you to set up a PPP connection with varying degrees of success. Now you really can get all you need to set up an internet in a CD. It is more a matter of deciding which part of the internet you want to be.
10. Multiplatform capability. Samba and netatalk provide cross platform capability for Windows and Mac
So I agree wholeheartedly that there is a business opportunity for Linux Admins. I will be ready someday soon!