Well, that is the situation we’re in. Go on with MT is cheapest option but definitely not the best. And in the present market it has no capital value since 90% of the market in Europe is not MT…
So if you don’t foresee a long future any longer, you have ample options:
Sell your customers. Since any new party needs to give them new CPE’s your customers by themselves are not very valuable.
Change the network yourself as long as you can finance it. Now you can go different options. Go for ubnt or eCambium which make your network more valuable but not any better in competing with others. So it has to happen fast and cash your network before its too late.
Or you can change your network and go for something new, like Mimosa. Time will tell if it makes your network valuable for sale, but it definitely will put you ahead of the competition so you can gain more customers at the expense of the competition and thus increase the capital value of your network.
But it has to happen within the next half a year. They rolling out LTE in Europe fast and we can only keep pace with their contracts if we can at least offer 25-30Mb without data limits etc. So we need high capacity networks. MT and ubnt will have problems in serving more than 15-20 CPE’s per AP with these speeds. That’s not going to do it…
I just removed the last 802.11a RIC unit two weeks ago. But we still have some AP’s running single chain with some 5% of my CPE’s still running single chain ‘n’ NV2. We just installed the first ‘ac’ P2P links with some netmetals and high gain antennas. And indeed, I also have 1 Netonix in use and 4 more in boxes ready to be installed. They are sooo nice…
Indeed we’d plan to start replacing only there were its needed first and where we can afford it.
Up to now our business plan is based on monthly surplus revenues only available for investments. Because of the crisis its almost impossible (and if it is very expensive..) to finance big time upgrades of hardware. So it goes little by little.
But we are now looking at leasing and even some dealers start to offer financial arrangements now. (This made eCambium very interesting). The idea now can be you replace some of the bigger AP networks at once. 2/3 weeks of hard work and done and after that you pay the next 2 or 3 year off what you installed. That’s opposite to what we do now.
Wifi in Europe and US and Asia is definitely changing. Some lessons I ‘tasted’ in the ASLAN Madrid spring 2015 is that Wifi as an end user solution for bringing internet to the remote (‘rural’) home and village is big business nowadays.
Technology is advancing and the average turnover in technology use in networks has a 4 year life time.
I started in 2003 with b/g went for a while to Horizontal polarization, then to ‘a’, further to ‘n’ and tdma (NV2) and now we have already some ‘ac’ units. But spectrum becomes more and more crowded. When I started we had only one or two ‘other’ small players in my region, and Wimax at small scale. Now we have at least 6 providers with in total over 40 AP’s apart from the first LTE AP by Vodafone and more to come, all in a valley of only 20km’s in surroundings…
At the same time subscription fees fell from € 50,- a month in 2003 to around 20€ where we offered 1Mb in 2003 which is now 10Mb at least!
The only positive to us in all this is that Wifi products and services become cheaper all the time, and are more capable. Backbone capacity I get now for less than 6/Mb symmetric. In 2002 I paid 450€/m for a 2Mb satellite line! Best I could get my hands on in these days were 4 coupled 8Mb ADSL lines! Now I can get Gb fibre if I want…
One of the major issues that seems to be bothering all WISPS and thus the hardware manufacturers is interference and getting the best out of the spectrum. Hence all these new developments. Several manufacturers and even so many approaches.
Also, in the beginning of this century some big names let the low end of the wifi market (‘last mile solutions’ they called it. like it is of lesser importance…) for what it was; ‘low benefit’. But now this part of the industry is probably becoming a bigger money maker, or at least of the same scale, as the high end and thus do we see some big players coming back into wifi… look to eCambium (=Motorola). And some other new players must have seen how fast ubn grew over the last years so they want a piece of the cake too. Hence we see more and more profesional players starting to eat away a bit of the market share that ubnt made over the last years. (honestly, their sales exploded..)
At the same time licensed LTE from old (the cell operators) or new (some new players do LTE only, for fixed connections only) are trying to get a piece of that same cake.
So, although the cake in some countries is still growing, there are more and more wanting to eat it. And the cake has to be more feature rich with bigger nutritious value in it… The next year will be very interesting to see what happens in (South-) Europe. And what happens with my network…
Well, lets see. I never bought with them because they are 5-10% more expensive than my house provider from Poland. And how come they can list something not even on the ubnt website nor seen anywhere else? But ok. One of the issues is that in the past I had poor experiences with ubnt. And every time again when I try something new from them I end up with a disappointment. Their sales talk, even in their official documentations, doesn’t always meet in real life production use. And last, but not least. Although ROS is at times troubled with bugs, I just happen to prefer it 1000% above a pure web based approach. If you have some IP conflict or otherwise no IP connectivity you’re lost… with mac winbox (or telnet) we can still correct things. A priceless advantage…
We’re still installing Mikrotik CPEs. When you finished replacing them with new gear, I’ll already recovered the investment.
Me too. After all, we still are getting new clients on board and with SXT’s we have a good, cheap CPE.
Years ago, if you install a customer, he remained with you for many years. Today, people change operator every two years on average.
So you need a CPE that carry out its function, and very economical, so you can quickly recover the investment.
This is not the future? Mimosa, Cambium, etc either. Future is fiber.
I agree on the investment recovery. That’s why the CPE stays our property. When the clients cancels we just take it back, revise it and use it again with a new client.
I don’t agree on the fiber. In rural area it is just too expensive. Even in many urbanised areas in Spain its not an easy (=cheap) solution since streets are usually paved in concrete, buildings and houses usually have their utilities (power, phone lines and TV cable etc.) coming ‘by air’. To deploy fibre here is expensive and you need permissions. To deploy you need to invest but still don’t know how many subscribers you will actually get… With wifi you need an AP and then each subscriber basically pays for his installation (direct or per monthly fee). Depending on what contract we have our investment recovered within 6-8 months.
Yeah, some people want 50+Mb download and streaming 3D UHDI TV receipt in the campo for € 20,- a month. Well sorry, that’s not possible unless he pays a couple of K’s to get his private fibre from nearest town to his house…
If I’d want the best, I’d go with PmP450 but… how can I amortize a 200$ CPE when a customer pays 15$/month in average?
Mikrotik gives us a powerful, stable, quite resistant to noise, and cheap CPE. It’s all that I need today
True. MT is relative stable and good and cheap. I would prefer it in any situation where interference is not such an big issue. But like I said before, we are really having issues with throughputs on both PtMP and P2P networks due interferences. We need better systems to fight it. Sync, better protocols, circulair polarization etc. There are some very interesting developments in the market. Its just a pity MT is not advancing in that respect…
The question I would ask is 90% of the domestic market prepared to pay an extra say 200% increase in CPE cost ?
There is another possible option in that WISP’s move away from 2.4/5.8 unlicensed bands to a licensed band and finally get clean air for 80/160++Mhz bandwidth currently required for higher throughput?
I wouldn’t class replacing AP’s + CPE’s as “update cycle” but would call it a major financial and network change,
Another consideration is that existing equipment may not co-exist with the new equipment, do you stop or proceed to change 100% of your wireless equipment on a busy site rather than in stages, I expect there will be a lot more of “wait and see” done by WISP’s?
No, but it is the operator that has to find a way of financing it. We still manage to charge the ‘light’ users for installation fee. (to cover a bit on the expense of the CPE) but many operators don’t charge anything anymore. Router wifi, antenna, installation, all for free and no contractual term… You can only do this if you have enough financial resources and go for the long term… Big players don’t need to make a profit, as long as they grow their bank is happy, their shareholders are happy and they just pump more money in their network to get bigger market share. And one day they either will be taken over for millions (and everybody happy, except maybe the client) or they are that big they take others (=growth again and again bank and shareholders are happy…)
I mean, we have a national player that is offering similar services as us, but no install costs, no hardware costs, no contract terms but still they pay 8 months worth of monthly subscription fee to the external installer. They probably only make money on a client after 2 years… But that doesn’t matter, the are growing and that is what financers want…
There is another possible option in that WISP’s move away from 2.4/5.8 unlicensed bands to a licensed band and finally get clean air for 80/160++Mhz bandwidth currently required for higher throughput?
There is not enough spectrum available. So price is high thus only nation wide (or international) players can afford that. Its cheaper to stay in the unlicensed bands. But it depends a bit per country I suppose. Here in Spain some (coastal) regions are really overcrowded with Wisps but going into the centre of the country it’s different. I asked Landatel where they make most of the sales and 90% is in coastal regions. Spain interior is still greatly a desert when it comes to WISPs.
My home country Holland is also different. 2-4 big national (international) players have 95% of the internet market. Only in remote areas some niche players are around. And they have the 5Ghz band almost only for themselves! But heho, initiatives are now made for financial grants and solutions to get even the most remote farms connected to fibre… The aim is 100% coverage of highspeed internet (+100mB!) for everybody… Here in Spain I don’t see that happening in the near future. Or it must be new wifi technology with new spectrum will become available that can do the same.. Not soon I believe…
I remember way back when “n” came out that the guys at mikrotik said they were working on being able to sync the ap’s to be able to reuse frequency. A soon to be released upgrade would allow card on the same routerboard (ie a 433) to sync across cards and later firmware would allow communication between routerboards for syncing. I remember I was excited about this.
Firmware came and went may times over and no syncing… What happened guys?
Wow… Thread is really off topic from “wheres the spectral scan, and when can we see it implemented” I hope spectral scan isnt going to be like the syncing… Teasing us then nothing happens.
Oh, and while I think about it… When will we be able to sync across APs?.. Answer from mikrotik either way would sure be nice.
haha… well, I hope to see some of the guys in Madrid next month. Will be asking about wifi in particular. I already asked normis to bring his boxing gloves. We have some sparring to do about their soft/firmware upgrade policy…
I hope lots of other read this tread too. The more users join in our point of view the more chance is MT might listen to us..
It might look like we are criticizing MT and pointing to other solutions but in the end the hope is MT will come up with better stuff in Wifi land… They don’t have to invent the wheel themselves always. Just look at what other are doing…
I am a Mikrotik fan but they need some critics at times to move…
One possibility is to use licensed frequency for backhauls. This frees a lot of spectrum and is cheaper than licensing PTMP Frequencies. Another option is to use GPS-Synced Backhauls to reuse frequency. I dont say a product name as my last post got deleted (forgot this is a vendor forum). Other option is to use filtering technic or antenna technic to get more separation and a higher frequency reuse.
We considered LTE in 3,5 (we use MiMAX there) but the only product we got offered is to expensive for what it does and needs a lot of infrastructure due to it’s LTE roots.
That will probably the next step for us, licenced backhauls. But with that other brand we spoke about we should be staying out of trouble a bit longer.
I use high gain antennas, even on short distances, for backhauls. Sometimes it needs to tune the radio down a lot.
I am also going to try 45º slant setup to stay even further out of interferences. And last, but not least, we make use of several building/houses and set some backhauls up in such a way they are ‘invisible’ from nearby radio’s.
And on very short link I also play with some triple chain links… more data bandwidth in the same frequency and channel width.
I also looked at LTE, not so much for me, but a competitor has plans of offering it here in the region. They might have some interest in my client base. But I have question marks hanging around in my head;
They really only want 1 or 2 towers in our region, covering some 5-10 thousand houses… All with one frequency? 30Mb for 20€. Lets say they get 1000 clients in a year? 20:1 usage ratio, 5Mb general daytime client usage. That still means they need to connect 200 clients at the same time and pump 1Gb to these. On average! Peak usage will be much higher. People already want streaming video. Now it is still a bit difficult to get that in the rural area. With LTE it should all be possible and thus people go for it and start to use it..
I have a 4G/LTE mobile from Orange. At times I am impressed, 30 to 35Mb downloads easy! And high uploads too. NLOS still gave me tests of 15-25Mb down. But at regular times, special in the summer with many holiday makers around, but in the same location, it dropped to only 2-4Mbps?
This is a pretty new tower in the region so not a lot of users yet. How is this going to work out when more and more users get 4G phones and start putting their movies directly online where others will look at them..? All streaming, yeah yeah, the sky is the limit. Well, G4/LTE can be good, but not that good…
Cel operators have data limits for their clients, this new company says they have not. They can’t do that since they are aiming at the fixed installations for rural wifi like we server now. If they really come with 30Mb downloads for 20€ they conquer the market fast… to collapse under the load soon after?
I don’t know, it looks promising, but still have my doubts…
Start with licensed at the central backhauls. You’ll love it. We now replace our first 366MBit FDX Link which runs for years without a single problem. We upgrade with 880MBit FDX and move the old link to a new location for the next few years. You forgot these links. They are there and they work.
For short hops 24GHz is an option. We wait for an unsaid company claiming to bring cheap 60GHz links this year. This is for up to 1km.
The manufacterer offering us LTE for sure make it work. But asking deep questions to them you’ll find a lot of promises for the future. A lot of their features are with firmware upgrades still to come. They do 64QAM downlink and lower on uplink. They may scale better due to more engineering and GPS Sync but they do not have a lot of bandwidth to share. You can buy an incredible amount of MT-APs for the price of one of their Sectors. So you sure offer better performance with more smaller sites. Microcells with high performance is the way to go.
This topic like many others has developed into a sort of classroom discussion of related items around the posted topic until
the teachers (MT) comes in to address the class, I hope I don’t get a 1000 lines for this
Well, I do hope that MT does comment on the posted topic… it’s been a while without word from them and again, it’s important to a lot of us I’d we are to consider continuing to use MT.
At the moment I’m trying to figure out if I want to use MT with 3.65ghz cards or use ubnt rockets… unfortunately the ubnt has more features.
Also just to put my two cents in… we only use 5ghz in our network for ptmp systems as licensed hardware (and licences) are too expensive. However, 90% of our back haul links are licenced or at least outside this range like 24ghz which gives us a lot of spectrum to play with in the 5ghz… 160mhz channels may be possible for a site with 4 sectors
Of course they havent, why solve something users are complaining about. they now make an sxt with a 2.4ghz for programming… Why cant they put another 5ghz n radio in for doing a spectral scan…
Its good that this tread is blown into new live again. It looks MT has started to work on wireless again so maybe by crying for it they might consider putting ac spectral scan, ac power output reading, nv2 background scan (or units with 2nd radio for scan) etc.
Wireless is at the moment the bottleneck in MT’s product portfolio. We need a lot of improvements or many of us will move to another platform over time…
We have now installed some Mimosa links, with quite decent results, but its just a dumb bridge. Mikrotik can do so much more. The spectral scan will be a major improvement already.
I’ve been look at them and was very enthiousiastic about them. Until I found their introduction of their AP’s was delayed by some months (into early 2016 they said for Europe. Haven’t heard anything yet…) because they needed to tackle issues they had on their backhaul units.
I also have been reading in some forums about dissapointing results on backhauls and actuall say some test somewhere where ubnt was outperforming them. Even a link with MT’s netmetals beat them…
But in theory their’s look very impressive. Especial their AP’s technology could implement loads of improvements on the PtMP performance, even in crowded environments. But so far not available yet (in Europe at least…) and to use it you need to swap your whole AP network.
So if MT could improve their PtMP performance (even faster chipset, MuMiMo, background scan in ac, frequency hopping, etc. etc.) than I’m sticking with MT.
Swapping a 50 client PtMP MT network for Mimosa would cost me some 8 to 10K (€) and a shit load of work. So if MT could make me the ultimate AP I don’t care if it cost me a couple of hundred euros… at least the clients can just swap like that…
