Ask Slashdot: Best Open Source Project For a Router/Wi-Fi Access Point? 193
An anonymous reader writes "My wireless router just died. I have an old netbook lying around that has a wired network interface and a wireless one. The wireless card is supported in master mode by Linux, FreeBSD, and OpenBSD. What does Slashdot recommend I use to turn it into a router/wireless access point? DD-WRT? pfSense? Smoothwall? Fedora/Ubuntu/OpenBSD with a manual configuration? I'm not afraid of getting my hands dirty and I know what I'm doing, but I want as close to zero maintenance as possible."
I've never actually used (Score:4, Interesting)
Re:OpenWRT (Score:4, Interesting)
Yeah. OpenWRT is the way to go. You can build or customise everything and it has pretty decent documentation in the wiki.
Development for DD-WRT is a mess and the documentation is awful. You can't trust the DD-WRT website with its database since it is massively out of date and has a ton of just plain incorrect information. You need to work out what specific blessed build number works for your particular hardware and revision but the only way to find that out is to trawl around in their forums with the huge threads. Once you've found a working build then don't ever upgrade since it is likely that they've broken it in mysterious ways in a later build.
Beat bufferbloat while routing (Score:4, Interesting)
Re:OpenWRT (Score:2, Interesting)
As somebody who spent about 3 weeks (I'm kinda new to linux) trying to get OpenWRT working on my router I would like to disagree. I can't speak to DDWRT's organization but the OpenWRT community seems completely dead to me: the wiki is outdated/inaccurate/contradictory (often on the same page) and the forum seems dead as well except from one or two threads. Good luck finding any help from that "community" *.
I was exploring DDWRT at one point and that documentation said OpenWRT packages will work with DDWRT. I don't know if that's true, partially true or untrue but I don't think I'm ever going to bother with OpenWRT again. If that is true of DDWRT at least for my purposes DDWRT will be just as flexible as OpenWRT. Also from what little I observed the DDWRT forums seem to have constant activity, the supported hardware list is much larger and the documentation much more complete/better written. In fact if you dig deep enough you'll find that OpenWRTs seeming officially supported hardware list is maintained by a completely different entity then that of the people in the forums and the forums is actually the place the look for your router.
* First I had to figure out my router would only run with the bleeding edge daily builds. Then I was trying to setup using local storage on router's USB port(s) then I was trying to get tftp-hpa configured, then I was trying to make the local storage/tftp daemon start/stop with a button press. Too much to ask I guess. Probably a little different if I had only wanted router functionality.
Re:DD-WRT on Buffalo hardware (Score:5, Interesting)
I'm a little bit surprised to see DD-WRT getting such prominent billing. I've been using OpenWRT very happily for a long time, and had trouble getting DD-WRT to do what I want. It's possible that things have changed since I last investigated, of course.
I'm a bit biased in that I wanted something hackable; I've been able to make packages for OpenWRT and have them work with very little effort, and even been able to debug stuff under gdb on the router. This is probably also possible with DD-WRT, but when I investigated, OpenWRT seemed clearly easier to develop on. Building the router image from source was dead easy; customizing it was easy with "make menuconfig" and building packages within the build tree (with support for the packages in "make menuconfig") was easily done as well.
My point here isn't to say "don't use DD-WRT," because I have nothing bad to say about it; rather it is that it's worth considering OpenWRT as well. Personally I've had a lot of success with it, and recommend it highly as a development router OS.
Hybrid; use a router as a wifi bridge (Score:4, Interesting)
Re:zero maintenance (Score:2, Interesting)
I had DD-WRT on my Linksys WRT54G v1.1, and it became flaky after I moved house (the wifi would intermittently stop working necessitating a reboot, not sure why but perhaps due to interference with a neighbours router as next door only switch theirs on when they are using it), but it became rock solid after I switched to Tomato (and I'm still using it).
Re:DD-WRT on Buffalo hardware (Score:4, Interesting)
Many folks make the same god damn mistake on the energy cost savings by getting a god damn router. Sorry but even in California (most expensive electricity in the U.S. @ 0.15 per kilowatt) your cost savings would need to be in the 100w per hour range to even be a Reasonable ROI (return on investment). $0.15 x 1kw (100 watts x 10 hours) is a meager $1.50. Take a cheap $40 router and you're looking at half a year minimum the payback to be worth it - that's for a system using 100 watts per hour. In the case of something low powered like an Atom based netbook - 20 watts for the CPU/Chipset (No Screen/GUI) you're now talking 2 years or better to even break even.
As an example, I have an old HP system with a 700Mhz Celeron (P3 era) using the Intel 810 chipset. Max power is 100w (PSU rating). In testing, the system uses 10w and can successfully boot from a floppy drive (pfsense). Note that it has 512M of ram (most $40 routers only have 4M) plus the CPU is 10-20 times faster. Hell pfsense or anyother floppy based router OS can damn near fit into the CPU cache. Now the cost to convert this to a router/ap is $0.00 for the simple reason it's already got a 10/100 nic and I have an unused b/g wifi card (PCI based) that can be installed. All I need is a bit of time to configure the unit and once done, it'll outperform any >$100 SoHo router on the market simply because of the amount of memory and cpu speed. All it will cost is some time/effort to setup. Value is priceless because I also learn how to do it.