Slashdot Mirror
Start Your Own Open Source-Based Telecom
prostoalex writes "George Ou shows how with the help of open-source VOIP server Asterisk you can start your own telecommunications company for under $6000 '...you can build a phone system that can support 72 analog telephones or fax machines, 100 IP hard or soft phones on site or remote, a T1 line to the public telco for 23 simultaneous external PSTN connections, multiple IP-based IAX trunks to multiple remote offices for seamless toll-bypass 4-digit dialing, IVR, and almost unlimited voice mail for everyone - for under $6,000 in a 1U chassis. Such a price point is easily 10 or more times cheaper than a commercial alternative,' writes George."
mah-Bell
An Indian-American Hindu committed to non-violent thought/speech/action alarmed by the global explosion of radical Islam
you discover that office class IP telephones are expensive as all hell.
we went with a NEC digital phone system with 2 wic cards for T1's incoming. the CSU cost us $12,000.00 but the Phones are only $185.00 each.
phones of the same quality in IP phones are neat $350.00 each, and that adds up fast when you look at around 100 phones plus 2 smaller CSU's that are set up as virtual offices at the ends of other T-1' for the sattelite offices with analog fallback if the connected T-1's fail plus allow us to bypass long distance charges by using least cost routing.
dont get me wrong, but an asterisk solution to replace what i just bought would be close in price and require a few weeks to get it working. I simply pay the local company to install it and maintain it.
Mod -1, Arrogant
here at work on a test machine, and at home on my XBOX! It's really incredible all the features that are available in Asterisk. Once you get the hang of the configs.. and there's probably only 3-4 configs you will mess with, it's a breeze. I have call routing rules setup to call my house phone, if I don't answer in so many seconds, it will dial out to my cell phone. Someone has also written a bluetooth presence script so it knows when your at your desk ( as long as the server is near your desk ) and when your gone so it knows which phone to call. Pretty slick. Not only is Asterisk been fun to play with, I've learned tons about telecom that I didn't know before.
It's nice to see such an impressive setup for such a low price, but to "start your own telco" for real, you'd need a bit more I think:
- Billing and invoicing software
- Provisions for wiretaps (if mandated by your local gov't)
- Customer service (unless you're not going to provide any)
etc.
If construction was anything like programming, an incorrectly fitted lock would bring down the entire building...
I work for a moderate sized telephony services provider, and I can attest that getting above the single T1 "toy" deployment written about here gets really, really hard. The Digium cards are crap, there's very little documentation, and if you try and run multiple carriers then have fun. In a few years this stuff might be pretty decent, and for small office deployments its great, but other than that it is ass.
We just did a VOIP implementation at work this last summer, and while we did not go with asterix, we did get a pretty good price (went with shoretel). However, the biggest single cost of all, was not the T1 installation, or the servers, or the software, it was replacing phones. Replacing 65 phones at our company at about $280 each was a pretty penny. Of course, with asterix you can find relatively cheap SIP phones, but they don't have all the fancy features, LCD's, that make it easy to use, POE, etc. So a hint for all you thinking of looking at VOIP, look at 3 very important things:
1. Yearly costs (maint, support, etc)
2. Upgrade costs. (how much is it to add each additional person. ie. phone+system capacity+licencing+anything else)
3. purchase costs ( for everything, T1 installation, installers, etc)
Look in that order. Many are cheap up front, but their phones are proprietary and cost a fortune, or can't expand in the chassis.
What are we going to do tonight Brain?
Check out the Sipura SPA-841 ($90. at voxilla.com, etc.). That is MUCH BETTER than the Fujitsu 'office-class' overpriced phones in common use in my 1000-line office.
I'm setting up Asterisk right now for use in our small office. Actually, it's basically setup now and I'm just waiting for my phones to get here.
;)
.. but then you'd have to dial it like an external number, and the other office wouldn't be able to call an extension directly).
Total cost for the hardware was under CDN$2000 (8 phones, 2-port fxs adapter for analog phones/fax machines, 4-port fxo card for incoming lines, and the PC). I probably spent about 40 hours total after deciding to use asterisk learning about it, configuring everything, and testing. Even at billable $60/hr, that works to $4400, which is a lot less than a comparable commercial system (I got quotes). It didn't actually cost that much anyways, since I don't get paid $60/hr.
We now have a phone system that has an IVR menu, pratically unlimited voicemail, and every other feature you'd expect in a phone system, plus when we open a branch office later this year I can use VoIP trunks to make intra-office calls pretty much free (and easy - encouraging communication between offices).
The system we have now is getting old, to add voicemail to it is $3000 by itself, plus the time to configure it (actually, I'd probably have to get someone to come in and set it up, since I only know the basics of how to program a few features). It can't do VoIP at all (unless you were to plug something into a CO port
This hasn't even gotten into the advanced stuff I can do fairly easily that wouldn't be possible with another system (without spend a LOT of money) -- such as, IVR status updates on system status; allowing customers to query their account balance etc.
Speak before you think
I've been looking at deploying a similar system in an office I work with. As far as I can see, the main advantage is thus:
I can take calls in, on a London number, have them handled by someone in the office, who puts them through to someone just like she normally does and they go to that employee at home. Provided he/she has broadband at home, no extra cost. Brilliant. It's not going to cost us $6000 to implement, more like $300/phone and $500 for the server, we don't need a system of the scale demonstrated, and should make the money back.
Not to mention, pretty cool!
Lemme get this straight: for just $6,000, I can start my own telecom, just like the big boys? I can bleed millions of dollars in red ink and employ mindless unionized drones who provide horrendous customer service while simultaneously driving my customers away? I can run miles of fiber through the neighborhood only to provide people with 384k upload speeds?
Wow, where do I sign up?
What's your damage, Heather?
This isn't a way to start your own telecom. There's no means of interfacing with the system at large other than by buying the services of an existing telecom at regular commercial rates. You can't, for example, realistically offer me and fifty of my random neighbors cheaper phone service in our houses with this. This is simply a way to build a PBX-type phone system that can inexpensively serve more than one physical location over an IP network. Timothy apparently doesn't understand the difference between being a telecommunications provider and simply owning a PBX or key system.
If a job's not worth doing, it's not worth doing right.
I have been doing just this very thing for the last year and a half. And let me tell you, it is NOT as easy as it would first seem. You really need a huge amount of capital to make this work.
I tried to self-bootstrap an asterisk based telco as my PRIMARY business supplimented with general Linux consulting with more than just $6k in the bank.
Here are some of the difficulties I have run into (and solved, but like I said, it has been a long, hard, expensive road):
1. The technology is COMPLICATED. This means inherently less reliable and big learning curve.
2. Asterisk is still unstable (even the stable version). A bug due to a completely untested patch added to the latest stable made me look like an idiot in front of a customer.
3. Standards are lacking. Asterisk often does not support all of the features of many voip phones. What do I tell a customer when there are buttons on their phones that don't do anything?
4. Asterisk has no billing system built in. Not a fault of asterisk but you can count on having to write your own. There are no existing open sources systems because they are everyones bread and butter. Nobody is giving theirs up so you can use it to compete against them.
5. Asterisk has no nice end user interface. Again, no real fault of asterisk but you can count on investing in LOTS of developer time. Asterisk configuration is complicated and to make an extensive interface is bound to be very costly.
6. I have had some bad luck with hardware from Digium. I am willing to chalk that up to bad luck. But the support from Digium is just unusable. I have left a dozen phone messages. Never once got a call back. I had to RMA a part that failed in production after just a few days of use. Yes, we tested the phone system etc and it all looked good. Then a daughterboard on the TDM400P (4 port FXO card) started causing the whole card to fail intermittantly. It took a lot of head scratching and days of calling digium, waiting on hold, eventually ending up in their voicemail box, leaving a message, and waiting for callbacks which never came to actually track down the cause of this intermittant problem. I originally started talking to them on Dec 17 regarding this. They suggested that the card was sharing interrupts and this was the reason it did not play well. On the 21st they said they had seen this problem before. On the 28th they admitted it was a hardware design flaw and offer to RMA the card. Why did they tell me to check shared interrupts then and waste a week of my time? I don't know. Around this time we find out that unloading the driver and reloading it would temporarily fix it but this had to be done on average twice a day. Note that the system is now in production. Worst possible case. So they are going to ship me a new card and I can send back the old card while we keep rebooting the system/reloading the driver on average twice a day. On the 29th very early in the AM I replied to their email with all of the info they need to ship me a new card and I expressed an extreme sense of urgency hoping the card would be overnighted the same day. On the 30th they emailed me an RMA number. I was told I could expect tracking info any minute. A couple days go by with no word from Digium. On January 4th I get an email telling me the card is on backorder! They expect more cards in on the 6th. So I on the 6th I email them to check if it had been shipped because I still had no tracking info and no card had arrived. This has all been interspersed with many phonecalls which were never returned btw. I am only citing emails because I have a record of them. On or about the 10th I call to see what the status is. The shipping personis not available but the operator promises he will call me back the same day wth some info. No phone call. On the 11th (yesterday, as I write this) I call again and explain I did not get a phone call. They are very apologetic and put me on hold while they look into it. After a few minutes I am informed that the card was never shipped! They promis