Any plans for the future of Mikrotik and the Broadband 2.0 world? ftth Is becoming abig kille ro f wireless but in our State here in New Mexico we have seen problems with companies like Qwest. They are trying to deliver broadband 2.0 but they can not. If Mikrotik had a project to spear head this market I am sure that this will solidify your place in the world of wireless choices.

I love my Mikrotik's but I need to be able to provide more to my customers. Is NSTREME the best way to do it or is there plans for greater speeds. (Broadband 2.0 is 10Mb or greater) Please tell me you guys are working around the clock for a great solution. If NSTREME is the way to go then how could one implement a MESH from end to end like this and do you have a recommended Routerboard for such projects?

-Sincerely,
DesertAdmin

i don't think Broadband 2.0 exists. What is that?

Sorry, mesh isn't comparable to fiber's capabilities.

It would be nice though to see MT offering Fibre Optic based alternatives as well. We are planning to deploy a moderate (5000+ clients) FTTH test roll out this year, and would definitely give MT a preference if they had something to offer.

Regards
HASSAN

http://www.wired.com/techbiz/it/news/2008/04/broadband2

This speaks of Broadband 2.0

Here is a company that did offer broadband 2.0 via wireless, but they are now shutdown due to a law suit with Qwest. I guess law suits with Qwest makes them angry and they don't like that so they will unplug you. Any ways. The company is called SkyWi which is wild because they reside in my state of new Mexico but they have customers in multiple states. So everyone with them are down Internet and VoIP wise. Which is sad. So while they are down we want to help their customers in our area. Sooooo.... We want to offer broadband 2.0 in the near future. Would at least doing NSTREME to the home be enough speeds for customers? What is the most you can get with NSTREME?

Oh here is that link of that broadband 2.0 company that does wireless.

http://skywi.com/

I also think that Mikrotiks can do this I just want Mikrotiks to be our solution they do everything that we need to have done.

Keep up the great work!

-Sincerely,
DesertAdmin

Thx for the links desertadmin, but as jp pointed out, Fibre is going to be the king of last-mile for the next 15 - 20 years, if not more. Fibre Optic has immense possibilities, and we're working hard to build a fibre optic network for the last mile in our country. Mikrotik has a great product, and all their equipment are excellently built and priced. I'm sure, if the put their minds to it, the could carry the same characteristics onwards to the Fibre Optic market. I hope the do so soon!

Thx for the links desertadmin, but as jp pointed out, Fibre is going to be the king of last-mile for the next 15 - 20 years, if not more. Fibre Optic has immense possibilities, and we're working hard to build a fibre optic network for the last mile in our country. Mikrotik has a great product, and all their equipment are excellently built and priced. I'm sure, if the put their minds to it, the could carry the same characteristics onwards to the Fibre Optic market. I hope the do so soon!

MikroTik ROS currently supports fiber interfaces. We're using an Intel Pro 1000 with LX fiber in one router and two 10G Fiber adapters in another. As long as the Ethernet chipset on the adapter is supported; it should work. As far as a RouterBoard with fiber, I'm sure they'll come. An RB1000 variant with a GBIC slot or two would be very useful (but probably also increase the cost).

Yes, dankerr. That is certainly possible, adding ethernet interfaces to PCs. But, what I would love to see is Fibre Optic offerings like RouterBoard. I would like to have CPEs from Mikrotik that have Fibre Optic interface on the Wan side. Granted it is going to be more expensive, but for wired networks who're doing FTTH using Active Optics, this would be a very good choice to move forward, given that you're having the power of the advancements in the Linux Kernel on each CPE. Wireless APs from MT already give this functionality to Wireless Carriers. We're still trying to mix this and that to get the same.

Yes FTTH is the best that is understandable. But for a WISP to survive in the market we must compete to a closer degree then b/g mode that is not bidirectional. We must become a bidirectional symmetric provider of 10+Mb(UP and DOWN) to the customer. The problem is that the CPE side I do not think can handle 10Mb+ symmetrically. Has any one else been able to do such a thing? We have an OC3 facility so we have the bandwidth we just need to provide it to the customers.

I am just curious to see what other people are doing out there and if it is possible to offer such a thing to compete with FTTH, cable and aDSL2.

-Sincerely,
DesertAdmin

We are not in a radio deployment at all, as we live in a very densely populated city (13M living on 200sq km). But, this year in Communicasia, I came across a company based in SG called SmartBridges who does seem to offer much power CPEs at quite reasonable costs. I did have a preliminary talk with them about having MT on the CPEs, which they don't advertise, as they compete directly with MT. The CPEs they make have quite a power radio inside, and they can do custom builds for you to allow for even higher output, within the regulatory guidelines in your area.

As for us, FTTH is the future plan, and we're still searching for stable equipment partners.

Broadband 2 is not a standard....

Broadband 2.0 is referring to providing more bandwidth to your customers. Virgin announced that they will be providing 50Mb via FTTH this summer in the UK. This is not about a standard for your country it is a bout a standard for the rest of the world. You can not live in a b/g world any more unless you can push more bandwidth like a MIMO 802.11N based cards. I am not talking about stringing fiber to rural communities. I am talking about being able to provide something that will allow us to stay in this industry and to give enterprise class solutions to our clients. If we do not then their will be no WiFi community. We must work now and at least try to push for more speed throughput. Those that post they can get what ever speeds always seem to post only UDP packet tests. Which is very idiotic. What networks mainly push UDP packets. hardly any most traffic is TCP/IP not TCP/UDP. I am just saying that we need to work together and explain where we are in the world and what is growing. As a WISP we must be competitive. Please band together and show more support for one another do not point fingers, we must improve and give valuable feedback to grow this product more.

Any ways....

NSTREME; no one has given me an answer on how many ppl are recommended per sector and what is the most throughput any one has seen reliably on an NSTREME customer to sector real world example.

Broadband 2.0 is a real term most people are just not aware of it yet. Lets make them aware by providing the speeds to our customers.

-Sincerely,
DesertAdmin

Broadband 2.0 is not a standard it is a term.

-Sincerely,
DesertAdmin

BroadBand 2.0 is a marketing trick word. Just provide EnourmousBand 999 to your customers and use whatever technology you and your economists team can calculate to be a good business for you. I would recommend digging for optics, that can be made cheaper when digging for gas+optic; optic for you+optic for other companies, etc. etc...

The Air can only take so much before becoming too noisy and noise bursts at times can cause service to quite simply drop all packets where optic fiber is rock solid. Competition will provide more via fiber like IPTV so get off your wireless asses and give reliable communications to your customers.

While you are correct in your statement that optics is the connection method of choice when it comes to maximizing speed, in many scenarios such is not an option because of exorbitant costs. In rural areas, where density is low, cost per customer for laying down fibre is very expensive. So, Wireless will be the preferred choice in such cases. I've heard of a new wireless technology which is currently being tested by NTT (Japan) called Super3G which can provide speeds upto 250Mbps per CPE. Granted this is more of a telecom technology, not broadband, but as days are going by, it appears that all last mile technologies are converging to provide the same thing --- bandwidth. What the customer does with that amount of BW available to them is becoming an entirely different ball game, with more and more content becoming available over the IP network. I won't be surprised if in a few years time all the famous TV channels start being offered over the internet, with HD.

Hmm fiber... Sorry that is just not a feasible option. Well does any one else have any suggestions I still think that NSTREME is the answer for now.

-Sincerely,
DesertAdmin

Cartes is very correct on the rural area deployment. That is pretty much the boat we are in. It is nor feasible for a fiber roll out. We are in a very NON-affluent part of New Mexico. So wireless is the only competition against Qwest's DSL. And we want to be able to provide more than just a few Mb but closer to 10-15 Mb this will secure our future and allow us to be a better provider. Our problem is that when we use NSTREME as a back haul we can only push about a 30Mb and even 24Mb to a client using 333 boards and 532RBs. I would love to hear a story of someone who is able to do such a task using the same type of boards. Are ppl using nstreme with turbo or not etc..

Any ways thanks for all of the replies.

-Sincerely,
DesertAdmin

How about nstreme-dual? With two Wi-Fi interfaces, with one dish with dual polarization feeder?

I may suggest you to look at Wimax 802.16d for such speeds.

Moreover, MIMO 802.11n may help aswell.

How about jumping directly to LTE equipment?

The magic that you are seeking is not going to come from Mikrotik. They just implement the available miniPCI radios and right now Atheros seems to be the leader in low cost chipset design for consumer radio cards. That is what our businesses are based on. The standard is 802.11a/b/g. a and g have a maximum 54Mbps radio throughput (about 36Mbps data) on 20MHz channels using 64qam modulation. You can squeeze the most out of this with faster processors (like the RB4xx family) or by bundling channels (Turbo mode or nstream2) but there will be no great improvements until they jump to the next level in modulation.

Right now we just have to keep our business focused and make money with the wireless products and technologies that are available to us.

Improvements will come when consumer devices jump to higher modulation rates, coding rates, and when new IEEE standards for the frequencies that we use are finally ratified and deployed. There are a lot of smart people out there working on this complex problem but don't expect the limitations to change much before the end of this decade.

Here is an interesting paper from 2005 that talks about this topic in detail.

http://w3.darkwhole.de/~tf/uni/mobilecomm.pdf

Tom

So only problem right now are the frequencies? We are still limited to 2.4 and 5 GHz ? But we can use LTE equipment that works on that frequencies, as long as a manufacturer would provide it for us. But if we want 1000+ devices, we could simply purchase it. Right?

So only problem right now are the frequencies? We are still limited to 2.4 and 5 GHz ? But we can use LTE equipment that works on that frequencies, as long as a manufacturer would provide it for us. But if we want 1000+ devices, we could simply purchase it. Right?

Uhhh right. Just not ready today. Figure 2012 for mass markets.

ISPs will probably be cut out of LTE. Or are you a cell carrier?

Tom

Yes, dankerr. That is certainly possible, adding ethernet interfaces to PCs. But, what I would love to see is Fibre Optic offerings like RouterBoard. I would like to have CPEs from Mikrotik that have Fibre Optic interface on the Wan side. Granted it is going to be more expensive, but for wired networks who're doing FTTH using Active Optics, this would be a very good choice to move forward, given that you're having the power of the advancements in the Linux Kernel on each CPE. Wireless APs from MT already give this functionality to Wireless Carriers. We're still trying to mix this and that to get the same.

Here in the US there are several ongoing tests of fiber into the home by telephone companies. A friend of mine is part of one of those tests. Several Cable companies have also ventured into this idea. The cable companies have backed off and started pushing DOCSIS technology along for home use (my current home connection is 30MbsRX/6MbsTX). The telephone companies have held back from widespread deployment of fiber to the home.

While both cable and telephone companies are using fiber heavily; they are not using it for the last mile. They use it for backbone and distribution; where it's then delivered to the customer over copper of some kind. In some cases the use it for delivery to a business. Why? My guess is maintenance costs; fiber is just not a durable medium for home use. In a business you can control access to the fragile fiber connections; in a home you can't. Every time someone pulls on their fiber too hard you have a truck roll, which costs money. If the baby stretches a coax or Ethernet cable it's not a big deal, stretch a fiber cable and the connection is gone.

There's also a big difference in cost between fiber connectors and coax connectors and RJ45's or RJ11's. In some instances fiber cable is cheaper than copper (coax, telephone or ethernet), the connectors and tools are not.

Even in a controlled data center environment; I see issues with fiber regularly. I've seen a spec of dust take down a 10G connection more than once. When you deploy something into a home; it needs to be durable. Long term, I don't see fiber as a CPE option; unless they find ways to make it much more durable.

Most of the TV over fiber is analog, unlike let say IPTV where the led/laser is putting out simple digital on/off. You can have digital TV over it, but they just take the RF signal off coax and convert it to light and send it over the fiber. The expensive CPE converts it back to RF for your TV tuner. This means the cable/telco has to deal with special connectors and low reflectivity connections and splices to maintain good SNR over fiber. Those companies don't want to go about television a different way and avoid all that special testing and engineering requirements.

Personally, I've only done Ethernet/Fast Ethernet/Gig Ethernet and T3 over fiber. It works good. It is very reliable. No lightning issues because fiber isn't a conductor.

So only problem right now are the frequencies? We are still limited to 2.4 and 5 GHz ? But we can use LTE equipment that works on that frequencies, as long as a manufacturer would provide it for us. But if we want 1000+ devices, we could simply purchase it. Right?

Uhhh right. Just not ready today. Figure 2012 for mass markets.

ISPs will probably be cut out of LTE. Or are you a cell carrier?

Tom

You don't need 'cells' to use the technology for point to point or point to multipoint. All you need it it runs on 2.4 or 5G so

1. You need it on allowed frequencies (2.4 and 5Ghz)
2. You need the speed

nothing else. you don't need cells you dont need GSM complexity, you need something much simpler.

So, how about a hardware manufacturer making those devices for you? License for (some of) the technology?

p.s. regarding the above post - even in Bulgaria companies offer fiber from several months and it costs as much as the Ethernet connection - cheap cheap cheap the technology is Passive Optical Network, look into it. And furthermore they can now supply High Definition IPTV

Personally, I've only done Ethernet/Fast Ethernet/Gig Ethernet and T3 over fiber. It works good. It is very reliable. No lightning issues because fiber isn't a conductor.

There's a big difference between reliable and durable. The company I work for depends on many fiber connections (both IP and other) and they are very reliable. The issue is durability; fiber is just not a rugged medium.

For widespread deployment in a "home"; the equipment and connections need to be durable and rugged. I've yet to see a fiber cable or connector that was durable and rugged.

Current technology is good enough for FTTH, search more and you'll see.

Current technology is good enough for FTTH, search more and you'll see.

I know it's out there:
http://www.ftthcouncil.org/
http://www.ftthcouncil.eu/

But "good enough" isn't a good idea if you're trying to build a viable business. Most of the big FTTH installs I've read about are done by a local government; who aren't as concerned with making a profit, let alone breaking even.

It still has durability issues, which is why most providers are going for fiber distribution with copper as the last link. It also has high maintenance/labor/support costs; which is why large providers like Comcast, Time Warner, Cox and the "phone company" are still using copper for the last link. Many have tried FTTH and have ended up backing off to FTTC (curb), FTTB(building) or FTTN(neighborhood). They want the fiber terminations where they have full control over access.

http://en.wikipedia.org/wiki/FTTX


My company has facility's in about 70 US cities and small towns. We've done much research on internet connectivity for each. We use fiber where possible; but many ISP's still prefer a copper hand-off even for a business. When asked why; they usually says it's that fiber requires more support. Some are willing to at least drop the hand off device (usually a media converter, but sometimes a router) in our space so it's easily accessible.

Here in Atlanta there are several ISP's doing fiber for business, most use a FTTB approach of some kind. We have multiple links from two of them, both layer-2 Vlan and layer-3 internet connections. Once is willing to do direct fiber connections (if your equipment is capable), the other prefers to drop a media converter and hand off as copper.

We're doing Fibre to the Business (FTTB) at the moment using active optics as opposed to passive optics. The last mile is also fibre, and usually the drop cable ends in a safe place of the customer's premises, from where the link is converted to RJ45 copper media before connecting to the customer's LAN. We're still not doing FTTH, but in my understanding, the same principle of carrying over the cable to a safe location inside the house, probably the cable box or something, and from there connect to the in-house cabling system. That would eliminate the cabling durability issues that dankerr is referring to.

As for the cable coming into the house, I've see fibre optic equipments made for Europe where they have a fibre holding tray inside them. Just a bigger casing, but makes it simpler to have the strong cable used for outdoor, to come into the house, and get into the box, fastened to the chassis. Inside the box, the core is spliced to a simple patch cable which in turn connects to the transceiver which is also inside the box (not outside, so no harmful laser sockets are exposed). So, no "accidents" from kids pulling the cable mistakenly.

From Mikrotik, I was looking for standard routerboards like the RB152 or RB453 or RB493 where the RJ45 ports are replaced entirely by Fibre Optic Tranceivers, which would make more sense for applications like us. Granted, they would be more expensive, but they would be at least cheaper than the alternatives that we see in the market today. I'm sure there would be ample demand for a product with 24 or 30 fibre optics ports, with the RouterOS loaded (like the 3rd party Powerouter), which will give active optics providers a powerful product to keep in their arsenal. And, the fibre optic interfaces are also becoming a standard these days, with SM Single Fibre WDM gaining ground every day. SC and LC is still an issue, but not a big deal as it affects only the patch cable.

We're thinking of rolling out FTTH in a big way in our country, and seems like we're out of luck in choosing MT extensively in the access network. So, we're looking for other vendors, some who claim to be providing a "manageable switch" with full fibre ports, which I wish I could do so much without if I had a similar offering from MT with high port density. MT Routers are so much fun! But perhaps I'm barking the wrong tree, as MT seems primarily geared toward the WISP market.

Thanx for everyone who've posted on this topic. Good to know what other "fellow ISPs" are doing across the world!

We're thinking of rolling out FTTH in a big way in our country, and seems like we're out of luck in choosing MT extensively in the access network. So, we're looking for other vendors, some who claim to be providing a "manageable switch" with full fibre ports, which I wish I could do so much without if I had a similar offering from MT with high port density. MT Routers are so much fun! But perhaps I'm barking the wrong tree, as MT seems primarily geared toward the WISP market.

Thanx for everyone who've posted on this topic. Good to know what other "fellow ISPs" are doing across the world!

I'm no longer part of an ISP or WISP (I've done both dialup and wireless), but I'm now one of the consumers... Having access to our co-lo facility and a good working relationship with each of our providers; I can tell you what some of their solutions are.

One ISP uses Cisco 6500 series chassis for core routing and core distribution. They then drop a Cisco chassis or layer 3 switch of the appropriate size in each building they serve. They then use various Metrobility devices as their demarc or CPE as needed.

Another uses Juniper equipment. And yet another uses Nortel equipment with a similar configuration.

There's also a new ISP on our floor that's using some new Cisco equipment that's geared specifically towards fiber based internet providers.

If you're looking for an inexpensive option (which none of the above are), you could do similar with MikroTik. While there are a few issues, ROS(3.17) does support Intel 10G Ethernet interfaces. You could use any layer2 or better fiber switch (depending on location in the network) and an x86 based MikroTik. Even Netgear makes some decent fiber switching.

From Mikrotik, I was looking for standard routerboards like the RB152 or RB453 or RB493 where the RJ45 ports are replaced entirely by Fibre Optic Tranceivers, which would make more sense for applications like us. Granted, they would be more expensive, but they would be at least cheaper than the alternatives that we see in the market today. I'm sure there would be ample demand for a product with 24 or 30 fibre optics ports, with the RouterOS loaded (like the 3rd party Powerouter), which will give active optics providers a powerful product to keep in their arsenal. And, the fibre optic interfaces are also becoming a standard these days, with SM Single Fibre WDM gaining ground every day. SC and LC is still an issue, but not a big deal as it affects only the patch cable.

We're thinking of rolling out FTTH in a big way in our country, and seems like we're out of luck in choosing MT extensively in the access network. So, we're looking for other vendors, some who claim to be providing a "manageable switch" with full fibre ports, which I wish I could do so much without if I had a similar offering from MT with high port density. MT Routers are so much fun! But perhaps I'm barking the wrong tree, as MT seems primarily geared toward the WISP market.

Thanx for everyone who've posted on this topic. Good to know what other "fellow ISPs" are doing across the world!

I've run fiber through the woods to some customers who are near some of our towers. I've also run it in our datacenter between our rack of wireless gear and our other racks of gear for the ultimate lightning protection. We also have fiber uplinks that I maintain.

We mostly use copper-fiber media converters to convert from the copper coming out of the MT or ethernet switch to fiber. These are available in the $50-60 range for either singlemode or multimode for 100 megabit.

It would be a nice option to reduce the amount of equipment at a site or have an real nice CPE by having a fiber port on a mikrotik RB. It might be nice for tower use as well for reducing lighting risk.

At the datacenter, I have found that by using not-the-latest-generations of HP procurve switches, there is a huge aftermarket of mini-gbic tranceivers that work well for cheap, and inexpensive single and multiport fiber modules for their modular switches.

If I were going to do a lot of fiber such as service provider work, I'd rather have a managed switch at the center rather than a mikrotik for its reporting and management capabilities. Currently an RB isn't a replacement for a nice managed ethernet switch. So I'd put in a switch for the customer ports, and vlan it all back into a powerful mikrotik for the internet access.

If you vlan it you will lose p2p bandwidth. I don't like tunnels and vlans. I like directly switched traffic as it takes the smallest route when its customer to customer.

If you vlan it you will lose p2p bandwidth. I don't like tunnels and vlans. I like directly switched traffic as it takes the smallest route when its customer to customer.

Vlans and Tunnels are extremely useful tools in networking. The packet overhead in both is minimal. I would guess by your statement that either you don't see any use for them in your network or you don't understand them. I suggest that you keep an open mind and study their application.

I am using vlans in some of my networks both to serve a purpose and to keep it simple.

Tom

http://forum.mikrotik.com/viewtopic.php?f=13&t=27716 Here is a WiMAX discussion that has very good information. The conclusion there is that MT should aim at 802.11n.

and here is the 802.11n discussion: http://forum.mikrotik.com/viewtopic.php ... 69&start=0

We have already done several fiber ring and fiber to the business deployments using some of our hardware as well as Routerboard systems. We use the core 2+ gig connections with distance fiber and then drop in smaller units like the 493s right at at the businesses!

MT will do this without issues!

Beyond 802.11n and Wimax, we have to aim at NLOS penetration just like Wimax. Wimax can do it at 2.5ghz why we cant at 2.4ghz ?

We are still using indoor protocols for outdoor LOS penetration. I think an open standrard has to be developed for outdoors just like Canopy. And this standrad should lead card producers like atheros to effectively produce the mpcis and other hardware stuff.

Doush has a very good point.

Why are we using indoor protocols for outdoor connections? With the newest release of 4.x the implementation of the 802.11n is very real for our network. This protocol primarily will allow us to handle many more protocol connections per CPE. I have also seen the competitior to MT have the 150Mb/s speeds to the CPE which is awesome but reliability I think still exists in the MT routerboard. The MT routerboard is advancing and I really hope to see higher Pthroughputs with some more of their development.

What is the most speed a CPE could get from an AP on 802.11n ( in a real world working environment?)
What is the distance at those rates?

Thanks for everyone's posts.

-Sincerely,
DesertAdmin