Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse).
Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection).
If required, we can have linux/bsd boxes we control at either end of the links.
If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)?
Thanks!
At layers 2/3 you can do Multilink PPP or even something wierd and questionably reliable like LACP over Ethernet-over-IP over the individual connections, but given that it's only 500m and you control both ends the best solution would likely be at http://www.netsys-direct.com/Ethernet_Extenders_s/1814.htm , particularly http://www.netsys-direct.com/product_p/nv-600ekit.htm
On Tue, Sep 29, 2015 at 4:18 AM, Trevor Cordes trevor@tecnopolis.ca wrote:
Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse).
Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection).
If required, we can have linux/bsd boxes we control at either end of the links.
If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)?
Thanks! _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
If you can put a box at both ends you can do compression between them with an arbitrary size lookup table that is dynamic based on the traffic.
To get the single TCP connection going over both you'll need to go up the stack - PPP or ipsec is what I would go with. You can't get away with a simple carp system unfortunately.
Either way I don't see how you don't have ecmp or similar over the lower link and run a tunnel with IP inside it.
Rob On Sep 29, 2015 7:10 AM, "Colin Stanners" cstanners@gmail.com wrote:
At layers 2/3 you can do Multilink PPP or even something wierd and questionably reliable like LACP over Ethernet-over-IP over the individual connections, but given that it's only 500m and you control both ends the best solution would likely be at http://www.netsys-direct.com/Ethernet_Extenders_s/1814.htm , particularly http://www.netsys-direct.com/product_p/nv-600ekit.htm
On Tue, Sep 29, 2015 at 4:18 AM, Trevor Cordes trevor@tecnopolis.ca wrote:
Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse).
Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection).
If required, we can have linux/bsd boxes we control at either end of the links.
If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)?
Thanks! _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
On 2015-09-29 A Myriad of MUUGers wrote:
If you can put a box at both ends you can do compression between them with
Yes, we fully control both ends, and can put a box at both. Compression could possibly help in our scenario too, as the traffic might be repetitive.
Far more effective, however, would be to upgrade to a symmetric VDSL2 setup that supports DSL bonded pairs.
Options that don't require replacing the DLSAM might be feasible if the cost isn't too high. VDSL looks like a possibility.
If you have Linux boxes at both ends, you can use mod_bonding in its round-robin mode... I've done that in the past and it does work.
That sounds promising too.
Are there any reasons for not considering commercial (backbone) wireless
We might do wireless. But the line of sight isn't great, and weather might be a concern? However, the application on top of all of this maintains one constant all-day TCP connection that can't tolerate being broken at all, and IME wireless is a lot less reliable than wires, even POTS wires.
Do you control the bare copper at each end? Adtran as a product called "ActivReach" which delivers up to 100Mbit POE over CAT3 style wiring.
Yes, we control the whole thing, including the copper. I need to find out exactly what cable is in there now to make sure it's CAT3.
We actually have 3 or 4 POTS runs to the 500m out location, and the DSLAM capacity is already there, and we have DSL modems too (so that's all "free"), so possibly trying a software-on-linux-boxen approach first would be worthwhile.
Many of the products y'all pointed out are really neat too, but if we need multiples of them it might start getting pricey, and from what I can tell would still require linux-boxen to aggregate.
Thanks for all the tips! MUUG shines once again. I'll keep you posted. I may have further questions as I implement. Maybe I'll do a presentation on it down the road (if it works out!).
I'm not clear on your topology, but the Adtran solution is a regular 1Gig switch with ActivReach capability. At the other end it's a dongle that converts back to regular ethernet. If you can put the switch in the middle it can service multiple sites so it might be cost effective. It also has a lot of advantages over any other kind of cat-3, or DSL aggregation. It's symmetric; full speed in both directions. Not a hack. And you can manage it just like any network switch so you can apply VLANs, QOS, etc. It's the only product like it in the market.
For anyone on this list that may need to deal with buildings where new wiring is a problem (asbestos is a typical example), this is a solution for utilizing existing "telephone" wiring to build a network.
If you are looking for carrier grade, then Adtran also makes DSLAMs with Gig over copper DSL type solutions, but I assume you are looking for cheap.
Regards,
John
On Tue, Sep 29, 2015 at 2:33 PM, Trevor Cordes trevor@tecnopolis.ca wrote:
On 2015-09-29 A Myriad of MUUGers wrote:
If you can put a box at both ends you can do compression between them with
Yes, we fully control both ends, and can put a box at both. Compression could possibly help in our scenario too, as the traffic might be repetitive.
Far more effective, however, would be to upgrade to a symmetric VDSL2 setup that supports DSL bonded pairs.
Options that don't require replacing the DLSAM might be feasible if the cost isn't too high. VDSL looks like a possibility.
If you have Linux boxes at both ends, you can use mod_bonding in its round-robin mode... I've done that in the past and it does work.
That sounds promising too.
Are there any reasons for not considering commercial (backbone) wireless
We might do wireless. But the line of sight isn't great, and weather might be a concern? However, the application on top of all of this maintains one constant all-day TCP connection that can't tolerate being broken at all, and IME wireless is a lot less reliable than wires, even POTS wires.
Do you control the bare copper at each end? Adtran as a product called "ActivReach" which delivers up to 100Mbit POE over CAT3 style wiring.
Yes, we control the whole thing, including the copper. I need to find out exactly what cable is in there now to make sure it's CAT3.
We actually have 3 or 4 POTS runs to the 500m out location, and the DSLAM capacity is already there, and we have DSL modems too (so that's all "free"), so possibly trying a software-on-linux-boxen approach first would be worthwhile.
Many of the products y'all pointed out are really neat too, but if we need multiples of them it might start getting pricey, and from what I can tell would still require linux-boxen to aggregate.
Thanks for all the tips! MUUG shines once again. I'll keep you posted. I may have further questions as I implement. Maybe I'll do a presentation on it down the road (if it works out!). _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
On 2015-09-29 John Lange wrote:
I'm not clear on your topology, but the Adtran solution is a regular
A DSLAM serves internet to a couple dozen computers on a small campus. There's a single endpoint at the far end of campus that needs to have LAN-type speeds back to the DSLAM for a specific application. It's just this one endpoint that needs this. This endpoint has 3 or 4 separate copper runs to it.
1Gig switch with ActivReach capability. At the other end it's a dongle that converts back to regular ethernet. If you can put the
Yes, the products people pointed out that enable this look very enticing. The Adtran one less so because it appears to only come in 28 port versions, and I can guess that the price would be >$1k. We'd only ever need 1-4 ports (for this one endpoint).
directions. Not a hack. And you can manage it just like any network switch so you can apply VLANs, QOS, etc. It's the only product like
Ah, but you're assuming we don't want "a hack" :-) Toying around with free surplus linux boxen to achieve what a multiple-thousand $ solution can do is right up my alley!
For anyone on this list that may need to deal with buildings where new wiring is a problem (asbestos is a typical example), this is a solution for utilizing existing "telephone" wiring to build a network.
Yes, I can see the value there when multiple drops are required.
If you are looking for carrier grade, then Adtran also makes DSLAMs with Gig over copper DSL type solutions, but I assume you are looking for cheap.
Ya, replacing the DSLAM at this point is not viable.
Aside: all of this begs the question, how are they getting gig out of crap cabling? What are they doing, technically? Where's Shannon's law? If it was always possible, how come this site is using a crappy slow DLSAM at all? Why should anyone install CAT6 ever when their junky CAT3 can do Gb? Confused.
Thanks!
Aside: all of this begs the question, how are they getting gig out of crap cabling? What are they doing, technically? Where's Shannon's law? If it was always possible, how come this site is using a crappy slow DLSAM at all?
On the carrier DSLAM side, the technology is G.FAST: https://en.wikipedia.org/wiki/G.fast . It's a relatively new standard and keep in mind it's still asymmetrical.
ActivReach is a proprietary Adtran technology. It's primary advantage over DSL-style technologies is it's symmetric, it behaves exactly like ethernet on the switch, and it's also POE. It works great to roll out VOIP phones in places where there is no other existing network cabling and re-cabling is expensive or not possible.
Why should anyone install CAT6 ever when their junky CAT3 can do Gb?
Confused.
Cost. You can get Gig from CAT3, but you can get 100Gig (or higher?) from CAT6 for the same price. ActivReach is not cheaper than re-cabling in an average situation.
Yes, hacks are fun and of course that's more or less what this list is about so I get that. But being the only person on earth that can support something that is probably not that stable and that people have come to rely on is maybe not as fun in the long term so it doesn't hurt to consider the commercial options. An "expensive" solution might actually be cheaper if you count all your time and/or pain.
John
This is an active area of research, particularly with the advent of multi-path TCP. Presently, however, you have to hide the two-link-ness from the TCP layer, and essentially from the IP layer as well. ECMP would work, as long as both lines are the same (this does not hold true as a dynamic assertion with DSL technology, *ever*). LACP will *not* work. If you have Linux boxes at both ends, you can use mod_bonding in its round-robin mode... I've done that in the past and it does work.
Far more effective, however, would be to upgrade to a symmetric VDSL2 setup that supports DSL bonded pairs. That'll set you back around $600+ per end, IIRC, replaces both the DSLAM and the DSLR, but makes your problems go away by turning all the copper into a single Ethernet link.
I just worked with someone else on this kind of setup, I'll see if I can find the links...
-Adam
On September 29, 2015 4:18:54 AM CDT, Trevor Cordes trevor@tecnopolis.ca wrote:
Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse).
Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection).
If required, we can have linux/bsd boxes we control at either end of the links.
If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)?
Thanks! _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
The unit I suggested someone else try was this: http://www.planet.com.tw/en/product/product.php?id=26560 Which can be purchased in north america via (among others) Amazon.ca: http://www.amazon.ca/PLANET-VC-231-Ethernet-VDSL2-Converter/dp/B009WWHNWU for only $142.83 ea. You need two; one for each end. No DSLAM required. This unit only does VDSL2+ profile 30a on a single telco cable at a time, and may or may not be able to support profile 30a at all. If 30a works for you, though, that's 100Mbps symmetric.
There's also BlackBox gear that's similar - the LinkGain extender (50Mbps, http://ca.blackbox.com/Store/Results.aspx/Networking/Extenders/LinkGain/n-42...) and the entire Ethernet-over-UTP series (up to 100Mbps, http://ca.blackbox.com/Store/Results.aspx/Networking/Extenders/Ethernet-over...).
Of course, you can combine two pairs of VC-231s with two linux boxes acting as bridges, each one configured for round-robin, and get an aggregate 200Mbps symmetric out of the cabling.
Good luck!
-Adam
On 15-09-29 08:09 AM, Adam Thompson wrote:
This is an active area of research, particularly with the advent of multi-path TCP. Presently, however, you have to hide the two-link-ness from the TCP layer, and essentially from the IP layer as well. ECMP would work, as long as both lines are the same (this does not hold true as a dynamic assertion with DSL technology, *ever*). LACP will *not* work. If you have Linux boxes at both ends, you can use mod_bonding in its round-robin mode... I've done that in the past and it does work.
Far more effective, however, would be to upgrade to a symmetric VDSL2 setup that supports DSL bonded pairs. That'll set you back around $600+ per end, IIRC, replaces both the DSLAM and the DSLR, but makes your problems go away by turning all the copper into a single Ethernet link.
I just worked with someone else on this kind of setup, I'll see if I can find the links...
-Adam
On September 29, 2015 4:18:54 AM CDT, Trevor Cordes trevor@tecnopolis.ca wrote:
Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse). Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection). If required, we can have linux/bsd boxes we control at either end of the links. If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)? Thanks! ------------------------------------------------------------------------ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
-- Sent from my Android device with K-9 Mail. Please excuse my brevity.
Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
To get a single TCP connection/thread going over two or more equal bandwidth connections you just need equal cost routing to use all paths. An actual router (not firewall, or nat gateway) with two equal cost routes will load balance packets over both connections as long as you're not firewalling. If you're trying to do this multi-connection extension at layer2 though instead of at layer3, you'll probably have a lot more hair-pulling fun, and less success.
Theo
On Sep 29, 2015, at 8:09 AM, Adam Thompson athompso@athompso.net wrote:
This is an active area of research, particularly with the advent of multi-path TCP. Presently, however, you have to hide the two-link-ness from the TCP layer, and essentially from the IP layer as well. ECMP would work, as long as both lines are the same (this does not hold true as a dynamic assertion with DSL technology, *ever*). LACP will *not* work. If you have Linux boxes at both ends, you can use mod_bonding in its round-robin mode... I've done that in the past and it does work.
Far more effective, however, would be to upgrade to a symmetric VDSL2 setup that supports DSL bonded pairs. That'll set you back around $600+ per end, IIRC, replaces both the DSLAM and the DSLR, but makes your problems go away by turning all the copper into a single Ethernet link.
I just worked with someone else on this kind of setup, I'll see if I can find the links...
-Adam
On September 29, 2015 4:18:54 AM CDT, Trevor Cordes trevor@tecnopolis.ca wrote: Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse).
Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection).
If required, we can have linux/bsd boxes we control at either end of the links.
If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)?
Thanks!
Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable http://www.muug.mb.ca/mailman/listinfo/roundtable
-- Sent from my Android device with K-9 Mail. Please excuse my brevity. _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
I disagree with Theo on this... thus illustrating that it's not a "solved" problem in the industry!
What I've encountered using ECMP for point-to-point connections over non-deterministic paths is packet-ordering problems so bad that I'm better off using only one of the two links in the first place! (Remember that one packet arriving too soon or too late nullifies the entire TCP in-flight window...) The following sentence is equally applicable to LACP as it is to ECMP, by the way...
WiFi links are the worst for bonding via ECMP, other types of wireless second-worst, DSL third-worst, and cable modems fourth-worst. Over plain old Ethernet, of course, ECMP works beautifully nearly all the time.
Then you have to deal with the fact that not all ECMP implementations are equal; the Linux kernel still either hashes the traffic (thus NOT giving you a 2x speed boost) or treats is more like an active/passive failover pair (again, no speed boost). As far as routers go, MikroTik appears to be the worst offender... it *looks* like it does full ECMP but it doesn't. Cisco and Juniper routers, naturally, do it right 99.999% of the time... at least you get /something/ for your money there.
Then you have to run a dynamic routing protocol to make use of ECMP, which drastically increases the solution complexity.
Pushing redundancy (and multi-path, and complexity in general) as far down the OSI stack as it can go has worked for me much better in all cases *EXCEPT* where it's impossible, and then pushing the multi-path capability right up to the application becomes necessary. Doing it in the middle... tends to break, in my experience.
-Adam
On 15-09-29 11:28 AM, Theodore Baschak wrote:
To get a single TCP connection/thread going over two or more equal bandwidth connections you just need equal cost routing to use all paths. An actual router (not firewall, or nat gateway) with two equal cost routes will load balance packets over both connections as long as you're not firewalling. If you're trying to do this multi-connection extension at layer2 though instead of at layer3, you'll probably have a lot more hair-pulling fun, and less success.
Theo
On Sep 29, 2015, at 8:09 AM, Adam Thompson <athompso@athompso.net mailto:athompso@athompso.net> wrote:
This is an active area of research, particularly with the advent of multi-path TCP. Presently, however, you have to hide the two-link-ness from the TCP layer, and essentially from the IP layer as well. ECMP would work, as long as both lines are the same (this does not hold true as a dynamic assertion with DSL technology, *ever*). LACP will *not* work. If you have Linux boxes at both ends, you can use mod_bonding in its round-robin mode... I've done that in the past and it does work.
Far more effective, however, would be to upgrade to a symmetric VDSL2 setup that supports DSL bonded pairs. That'll set you back around $600+ per end, IIRC, replaces both the DSLAM and the DSLR, but makes your problems go away by turning all the copper into a single Ethernet link.
I just worked with someone else on this kind of setup, I'll see if I can find the links...
-Adam
On September 29, 2015 4:18:54 AM CDT, Trevor Cordes <trevor@tecnopolis.ca mailto:trevor@tecnopolis.ca> wrote:
Is it possible to aggregate DSL lines, to combine them to get X-times the bandwidth on a single link? In this situation, I control both ends, the DSLAM and the DSL modem side on the other end of some POTS runs (CAT3-ish I assume, or worse). Note, I don't want load balancing or fancy routing/sharing. I need double (or more) the bandwidth for a single application (single TCP connection). If required, we can have linux/bsd boxes we control at either end of the links. If it's not possible, does anyone have any other ideas for somehow getting better bandwidth out of 500m POTS wires (quantity 4)? Thanks! ------------------------------------------------------------------------ Roundtable mailing list Roundtable@muug.mb.ca <mailto:Roundtable@muug.mb.ca> http://www.muug.mb.ca/mailman/listinfo/roundtable
-- Sent from my Android device with K-9 Mail. Please excuse my brevity. _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca mailto:Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable