Gentoo Archives: gentoo-user

From: Mick <michaelkintzios@×××××.com>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] To all IPv6-slackers among the Gentoo community
Date: Wed, 27 Nov 2019 09:29:24
Message-Id: 1686616.khKhjDU9G2@localhost
In Reply to: Re: [gentoo-user] To all IPv6-slackers among the Gentoo community by Dale
1 On Wednesday, 27 November 2019 01:51:44 GMT Dale wrote:
2 > Rich Freeman wrote:
3 > > On Tue, Nov 26, 2019 at 8:10 PM Dale <rdalek1967@×××××.com> wrote:
4 > >> I went to Newegg. Hey, I buy stuff there sometimes. Anyway, I've
5 > >> looked at several routers and none of them mention IPv6 that I can
6 > >> find. I even skimmed the reviews and can't find a mention of it. Is
7 > >> there some secret way to know when IPv6 is supported? Is it called
8 > >> something else maybe?
9
10 It is called the OEM's website where technical specs are provided for each
11 model.
12
13 It is also called Wikipedia. There may be a page where all models of a
14 particular manufacturer are listed in some table, explaining their
15 functionality.
16
17 There are also webpages with reviews - but careful with those. Most are
18 nothing more than a shill for Amazon or some such shop, rather than an honest
19 technical appraisal. Sometimes youtube may have an unwrapping video, or demo
20 of configuring a particular router - if you are interested to know what they
21 look like in more detail.
22
23 There are specialist websites like:
24
25 https://www.smallnetbuilder.com/
26
27 as well as open source firmware projects like OpenWRT/DD-WRT etc., with useful
28 blogs and forums to peruse, along with reports for suitable hardware.
29
30 Online shops are the last place to visit, *after* you have concluded which
31 router is best for you, to see if you can afford the price. Their websites
32 may have incorrect technical information, out of date specifications and
33 irrelevant (annoying) marketing speak to attract consumers. Many are just
34 box-shifters and wouldn't be able to tell you what's in the box you ordered
35 anyway: "errm ... whatever they're shipping from China these days". o_O
36
37
38 > > IMO there are three reasonable approaches you can take towards getting
39 > > a router you won't curse yourself for buying a year from now:
40 > >
41 > > 1. DIY. PC or other general-purpose computing hardware with multiple
42 > > NICs. There are SBCs that work well for this. You can run pfsense or
43 > > some other router-oriented distro/software/wrappers. Or you can just
44 > > roll your own with netfilter and such. Max flexibility, but also max
45 > > fuss. Unless you use a SBC you'll also be paying a price in
46 > > electricity. Don't underestimate how much you pay for any x86-based
47 > > system that runs 24x7 - especially anything old you have lying around.
48 >
49 > I remember how my old rig pulled power. It pulled like 400 watts or so
50 > idle. Of course, it was lacking in power so when compiling, there
51 > wasn't a lot of difference really. In the winter, I rarely needed a
52 > heater. Its constant heat output kept this bedroom comfy. No real need
53 > for a heater. It's one reason I want to avoid this option. Mostly, I
54 > want something I'll get many years of service from and everything work
55 > well, wired or wireless now that I have a cell phone and printer that
56 > needs it. My current router pulls like 10 watts or something.
57 > Considering I run electric heat and such, it's a rounding error for me.
58 > Heck, my main puter is too. It pulls like 180 watts which includes
59 > everything, modem, router, monitor and the rig itself.
60 >
61 > The positive part tho for option 1, if another port is needed, just add
62 > a network card and it's done. With DHCP and friends, it will likely
63 > just work. That's something you can't do with a store bought router.
64 > Whatever it comes with, that's what you got. I've never needed more
65 > than the 4 most come with tho. My puter uses one, printer another and
66 > cell phone. I guess I have one left still.
67
68 Every additional network card (PCI or USB) will also incur additional cost and
69 soon you could run out of MoBo slots. It makes more sense to buy a dedicated
70 switch instead, with as many ports as you think you will need to use in the
71 future. Old routers can also be used as dumb switches, after you disable
72 their DNS/DHCP/PPP, etc. The world is moving towards high speed wireless
73 connectivity anyway, so more and more devices will not need a physical switch
74 port or ethernet cables to gain access to the network.
75
76
77 > > 2. OpenWRT/DD-WRT/etc. Again it is a bit fussy but generally way
78 > > less so than going pure DIY unless you're running pfsense or some
79 > > other appliance-oriented distro. If you go this route then definitely
80 > > check for recommendations on hardware that is known to work WELL.
81 > > Some stuff technically works but can be very prone to having to play
82 > > around with JTAG and such if you make the slightest mistake. You'll
83 > > probably spend an extra $20 on hardware you won't regret buying - do
84 > > it.
85
86 +1
87
88 Trying to save a few pennies could result in being lumbered with suboptimal
89 hardware.
90
91
92 > That's what I'm wanting as a option. I may just use the firmware that
93 > comes with the thing for a good while. Later on tho, if needed, I may
94 > switch to Openwrt or some other option that may work better. It's a
95 > option I'd like to have if possible.
96
97 This is generally a good option because OEMs hardly ever bother upgrading
98 their initial firmware these days. They're more interested to ship the next
99 model, or the same model in a shinier box. As a result loads of routers are
100 running around the world with actively exploited vulnerabilities. At least
101 with OpenWRT and friends you have a live opensource project actively trying to
102 keep on top of the latest SNAFU.
103
104
105 > > 3. Something commercial that isn't terrible. There are various
106 > > options, but everybody always points to Ubiquiti and I'm mostly happy
107 > > with them. If you want something that is more gui-based I'd go with
108 > > their Unifi line. I'd avoid Amplifi as it is more consumer-oriented
109 > > and you'll end up being frustrated with it. EdgeOS is getting closer
110 > > to something like OpenWRT - it runs linux and you can get a shell and
111 > > mess around with the CLI. However, while the EdgeOS routing options
112 > > are great they aren't so good with WiFi and EdgeOS and Unifi don't
113 > > interoperate all that well (not impossible, but they don't really talk
114 > > to each other so you have to maintain two configs). I also really
115 > > dislike that the EdgeOS management software is only supplied as a
116 > > docker image, which is a pain if you're not using docker (one of these
117 > > days I'll have to get it working with my bridge interface as it always
118 > > tries to create its own and ends up having no physical network
119 > > access). The Unifi controller software is packaged for a couple of
120 > > distros which makes it much more flexible to deploy (and you can use
121 > > it on docker if you wish).
122 > >
123 > > Personally I'm running EdgeOS on my router and Unifi on everything
124 > > else. If I could go back I might have gone with Unifi on the gateway
125 > > but it does bug me that it is so much more expensive and does the same
126 > > thing. If I had it then end-to-end VLAN/etc would be much more
127 > > practical, though I'd need a pile of managed switches to make it work
128 > > well.
129 > >
130 > > I've run all three options at various points. Unless your needs are
131 > > special I think there is value in just going with #3. It just runs
132 > > itself for the most part, and if you want multiple access points or
133 > > anything like that the network basically runs itself. I just plug in
134 > > new hardware and then on the controller software it shows up, and one
135 > > click provisions it which configures it to fit in with all my global
136 > > settings.
137 >
138 > This is why I might buy one compatible with Openwrt but wait until the
139 > wireless stuff gets sorted out. Like I said above, I'd like it as a
140 > option so finding one that Openwrt supports should increase my odds if
141 > they get everything working nicely later on. I still remember the old
142 > USB days when it was new. It was buggy and stuff didn't work right
143 > every time. After a while tho, they got most the kinks worked out. I
144 > think Openwrt and others will do the same. It may take a bit but maybe
145 > by the time I'm ready to try it, it will be awesomeness.
146 >
147 > I just want to avoid replacing my current router with a router that also
148 > doesn't have IPv6 support and has limited options later on. Even google
149 > isn't helping me much on that.
150
151 Look at my suggestions above on how to investigate the availability of IPv6 or
152 other desired functionality of candidate routers.
153
154 Something I hadn't mentioned, merely because I don't know if it will work with
155 your old router, is to hack the hardware itself. Replacing the flash disk and
156 RAM with larger components may land you a more capable device for no/little
157 extra cost. Just use one of the RAM modules you have lying around in your
158 spares bin and hope the chipset is capable of booting and utilising it. Some
159 SoCs are crippled by design, having a max RAM capacity they will initialise
160 hardcoded in their boot code. They may not see or use more RAM and may even
161 refuse to boot with it. Nevertheless, it could be an interesting project for
162 a rainy day, on a router which is on its way out anyway:
163
164 http://neophob.com/2006/01/wrt54g-ram-upgrade/
165
166 --
167 Regards,
168
169 Mick

Attachments

File name MIME type
signature.asc application/pgp-signature

Replies

Subject Author
Re: [gentoo-user] To all IPv6-slackers among the Gentoo community Dale <rdalek1967@×××××.com>
Re: [gentoo-user] To all IPv6-slackers among the Gentoo community Walter Dnes <waltdnes@××××××××.org>
Re: [gentoo-user] To all IPv6-slackers among the Gentoo community Wols Lists <antlists@××××××××××××.uk>