Cisco VoIP Ditched for Open-Source Asterisk

An anonymous reader writes "Sam Houston State University (SHSU) is moving 6,000 users off a Cisco VoIP platform to an open-source VoIP network based on Asterisk. One big driver, of course, is cost. From the article: 'We thought that it will be more cost effective in the long run to go with an open source solution, because of the massive amounts of licensing fees required to keep the Cisco CallManager network up and running,' says Aaron Daniel, senior voice analyst at SHSU."

On the subject of Asterisk

I've just released FreePBX 2.1.2, which is a major security upgrade from 2.1.1. Not really relevant to this article, except that they both deal with Asterisk.

(For those that don't know, FreePBX is the only open source GUI for configuration and management of Asterisk. www.freepbx.org [freepbx.org])

--Rob

SCCP support?

Seems like the majority of Asterisk support has been for SIP phones. Some support for SCCP phones such as the 7910. Be nice if more low end phone support was available. Overall, Asterisk seems much nice than CCM and does not rely on a OS/Application installation.

Re:SCCP support?

I was really trying hard not to reply to _every_ post here, but SCCP is an awful protocol. And the 'low end' VoIP phone are all SIP or IAX, so you're barking up the wrong tree a bit. For example - Google for PA1688. This is a VoIP phone _chipset_ that the manufacturers have open sourced the firmware for. You can usually buy PA1688 based phones for about US$50. Or if you want more of an office phone, the Grandstream GXP2000 has a reasonably professional look, and are around US$100 or so. Going up market from there, you're looking at the Snom 320 or 360. Plenty of buttons and lights, and it runs Linux.

--Rob

Asterisk really is best bang/buck

I consult for a small Asterisk host, Lylix.net, and our customers couldn't be happier. It's a bitch to configure (hence we can charge $$$ for the service) but I'll be damned if it isn't a solid piece of FOSS, much like Apache. My hats are off to the Asterisk guys, it's likely to become one of the most important FOSS projects in the next 5 years or so.

Re:Asterisk really is best bang/buck

Actually, Asterisk isn't _really_ FOSS, as you have to sign a disclaimer (before you submit code to them) giving them the right to repackage it in non a FOSS way. This is so they can sell the Asterisk Binary Edition, as well as (unclear, to me) licencing issues with Intel Dialogic cards.

OpenPBX.org (nothing to do with my FreePBX project, mentioned above) is a pure GPL fork of asterisk from about a year ago, that they've done significant amounts of re-writing on, including working on a new dialplan language, as well as being able to import a lot of Steve Underwoods work (www.soft-switch.org) with software DSP (eg, soft-faxing, T.38 [fax-over-IP], better DTMF detection) that he will only licence under the pure GPL.

--Rob

Re:Asterisk really is best bang/buck

Not sure about mysql and trolltech (I think they are mostly developed in house, actually), but Apache uses the Apache license which allows for non-free distribution of the code. The contributors have to license their contributions properly to get them accepted into the main code base, but they don't have to give up their ownership rights.

Re:Asterisk really is best bang/buck

You seem to have a good knowledge of Asterisk, yet I have to correct you on the fact that Asterisk *IS* F/OSS and *IS* released under the GPL. What you're talking about is giving your copyleft to Digium if you want *YOUR CODE* to become part of the official distribution. Nothing new here, it's a common practice, used even by the FSF which *MAY* change the license then, but you can be pretty sure that the FSF won't change it to a non-copyleft license (while Digium uses it to give non-free licenses), but how do you think they'll change all code from GPL 2 to GPL 3 [not counting GPL 2 or later, since some of the GPL'ed software owned by the FSF (ie you give them your copyleft) hadn't the "GPL 2 or Later" clause and they added it later, since the license can only be changed by an agreement of all the copyleft holders, so it's easier if it's a moral entity like the FSF, MySQL AB, Trolltech the Apache Software Foundation (even if they don't use GPL, they still may want to change their license)... or Digium. And they all ask for copyleft transfer.

My point being: yes, Asterisk is "100%" F/OSS. They just don't allow other copyleft holders in THEIR distribution. Nothing would prevent OpenPBX, to sync with each latest version of Asterisk, but as long as Digium wants to hold all copylefts, they can't include code made by OpenPBX folks. Digium wanting to hold all copylefts is a part of their business model (dual-licensing). Of course, it makes it harder for OpenPBX people to sync because of the two development trees (and I understand why they'd want to keep their copyleft). However, Asterisk remains Free Software. Maybe they're not using the "Open Source development model" at its maximum though, but who cares :). As long as it's Free (with a capital F), it's fine with me.

Asterisk versus CCM features

From the article:

"While Asterisk and the SIP protocol lack some of the more extensive features on the Cisco CallManager..."

This may be true for vanilla Asterisk, but there is an extensive community adding a wide range of additional features and services to Asterisk. For example, <plug>our Enswitch product [integrics.com]</plug> provides a layer of billing and commercial services on top of Asterisk and SIP Express Router. Having work extensively with both Asterisk and CCM, I would claim that with Asterisk plus all the applications that work with it already surpasses the features of CCM, and Asterisk has the momentum behind it. Over the next few years, CCM will fall further behind, and before long Asterisk will be the dominant telephony platform in the same way Apache is the dominant web server platform now.

Asterisk in the workplace

My company sells Asterisk solutions to business clients and we're very happy with it. Once you figure out what you're doing the sky is the limit when it comes to configuration. My only issue with Asterisk is the voicemail subsystem. If Digium would put some time into that I would be the happiest person alive. Tom

Unversites are overrated.

This is not an attempt to troll or anything. But this doesn't seem like to me as a major blow to Cisco. Universities and Corporate and Government user are a much larger sectors at large compared to universities. And dont tell the College recruiters this the rest of the world doesn't follow what universities do. for the following reasons.

Universities have cheap skilled labor. A slew of talented kids/young adults who are willing towork for free or near minimum wage, but when they leave to the real world they will be demanding $35,000 and up a year for the same job. This is the reason why many Open Source projects work and save money in Universities but when a Corporation gets it, it becomes a money pot. Because for a company it is cheaper to call Cisco and pay them $1000 for a fix to their problems then having a team of 10 people at your company taking a day to fix the problem because they do not have the answer sitting right in front of them or able to contact the engineer who created it. vs. a University where this 10 people 8 bucks an hour are much cheaper then calling Cisco for help.

Universities are allowed to experiment almost by charter. If something goes wrong this screw all the people who are not getting phone service. You will have wait until we fix the problem, it is not like we are loosing money with the phones down for a couple of hours. Private companies loose money when their communication are done so they want Cisco to come and fix it right away and they better know what they are doing. Being an Education facility it is allowed to experiment in different products while Companies find better value in using what they know works.

Liberal University vs. Conservative Corporations, basically means if it not exactly what we want we keep on trying and trying until we get it right (perhaps making it worse in the process) or If it does what we need we hold on to it until we find the perfect solution (which guarantees that they are going to use a product they don't like for a long time)

This is why Open Source is popular in Universities but in Corporate and government use they need to work a little harder to get acceptance.

Asterisk needs improvement.

I work for a SIP hardware provider. We have a whole department dedicated to interoperability testing with other vendors of SIP infrastructure and user agents. Asterisk is approximately the least SIP compliant bit of software out there. It's great if all you want to do is basic calls but the reason why it's perceived as working so well is because vendors (like us) have to hack our software to work with it because our customers demand it, even if it makes us non-RFC compliant. Why has Asterisk never shown up at a Sipit bakeoff despite having been repeatedly invited? Asterisk has unfortunate momentum.

Re:Asterisk needs improvement.

That's totally incorrect. OEJ (a leading developer) has taken Asterisk several times to SIP Interoperability Testing meetings, and has acted very proactively to fix perceived or real incompatibilities.

I just did a quick search of the Digium bugtracker, and I didn't see any 'SIP Incompatibilty' bugs there apart from an issue with sipgate.de.

I honestly think you're trolling, or you have no concept of how FOSS works. If there's a bug, you fix it, and if you can't fix it, you report it and someone who can fix it, will.

--Rob

No PCWorld?

I mean, before bashing on other people, they might just have looked at their own site.. The page contains 160 images (is that really neccesary?)), When I hit back I got a "malformed URL" message, and frankly, it's just an ugly and awkward site IMO.

SCCP = Skinny?

OK I give up why is SCCP called "Skinny" instead of "Skippy"?

Spoofing Caller ID never gets old

My friend at my last job wrote his own VOIP software and he told me it wasn't very hard. If I had my own voip system, all I'd be doing would be making calls to my friends from the White House, Pentagon, or other famous places for fun.

Why do they price themselves out of the market?

Or, more puzzlingly (that a word?), how do some companies get away with competing against FOSS products with highly expensive proprietary offerings? I'm assuming that the proprietary solution has the same functionality as the other; maybe some bells and whistles on the fringes, but essentially the same.

They must make their money from licencing fees (and maintenance, but FOSS can do that, too). So why don't customers choose the cheaper option. Don't get me wrong; while I approve of FOSS and use it whenever I can, I won't hestitate to buy a proprietary product if it does what I need and there isn't a viable FOSS alternative.

I'm no expert in this - which is why I'm puzzled. Can anyone tell me (us) why? Is it any combination of the following?

1. "Noone was ever fired for buying IBM" (MS/Cisco/etc).
2. The bells and whistles are what the buyer craves.
3. Proprietary products have better support.
4. It's free, so it can't be worth anything.
5. What's FOSS?
6. We only run Windows (Solaris, whatever).
7. Proprietary products are better "rounded" or "easier to use".

I know that all these have flaws and, sometimes the reason is valid. But overall, I think my question still stands.

BTW. If anyone can think of anything to add to the list - I'd love to hear it.

Another open source alternative

I've used asterisk quite a bit and it works quite well. Also Sipx PBX is another good performer, although slightly harder to set up, easeier to configure. Sipx PBX is another open source solution that can be found over at the Sip Foundry [sipfoundry.org]. They have some good testing code that comes in handy when troubleshooting sip to sip issues. cluge

Asterisk?

I know everyone hypes Asterisk and Open Source and all that.

But has anyone looked at Asterisk close enough? It's the most horrid piece of software I have seen in a long time. Its configuration is awkward at best and downright inconsistent and nonsensical at worst.

Its documentation is practially non-existent. Nowhere do you find a good documentation written by the programmers. All you have are Wikis and web sites where people try and guess how Asterisk works. Howtos consist of config snippets without explaining what the options mean, let alone explaining the grand scheme behind everything.

Maybe it works after you configured it based on some other guy's experience, but if you want clean and well-documented software, go look elsewhere.

Asterisk seems to be the PHP or MySQL of the PBX world.

</rant>

Freeloaders or open source pioneers ?

There is some irony to this story - the expensive part of any phone system is (hold your breath) the phones. I will point out that the SHSU could pick an open standard protocol and move the phones from one system to another. Try that with Microsoft Office Communicator some time - you can't. I noticed that this story is under the Linux category and - I will point out that Cisco Call Manager 5.0 runs on linux and can run SIP to phones (as well as many other protocols).

Now, I know Asterix fairly well, Cisco fairly well, open source VoIP fairly well (as the joke goes I wrote the O'Reilly book), and SIP really really well. As was pointed out in Mark Spencer's Keynote at VON last week, the SIP stack in Asterix certainly has some room for improvement. And given SHSU does not seem to have any intention to support the development of Asterix by buying a support contract from Digium, I sure hope they are doing something to make sure that Asterix get the support that they will need it to have to stay relevant.

asterisk kills call manager

I work for a Fonality PBXtra reseller and the pbx absolutely rules. Asterisk on linux is the future of PBXs. The menu system, reporting, call queues and gui absolutely kill traditional phone systems. BTW, Vonage runs on Asterisk and so does broadvoice and other VOIP companies.

I'd say its a huge mistake

At my work we have Asterisk PBX/VoIP and its still way too raw to use. Unending list of problems and bugs, things work differently with new releases and it is not that stable or efficient under heavy loads. We end up having to purchase very beefy hardware and hire a guy whoes job it to monitor/fix it 24/7.

Asterisk-based 200 wireless/wired phone deployment

I just deployed an Asterisk phone system powering ~140 wired Polycom phones and ~70 wireless phones covering 31 acres. Here are some tips from what I learned in this process:

1. Pick a capable vendor for each job you outsource. I looked at Asterisk and decided it is too technical for a Asterisk newbie to build a production system, so I called Digium and they referred me to a dCAP certified Asterisk consultant in my area. Knowing Asterisk is one thing, but knowing how to pull off a great install is more than that. Our vendor developed a workbook that covers many parts of a successful deployment, such as reviewing the network (gear, configs, wiring plant), getting the users (names, current extentions, locations . .), getting the users to think about the dial plan and having them understand their satisfaction with the results is directly related to trying to get it right. When we distributed the phones to each desk, the boxes were labeled and sorted on the pallet this helped save a huge amount of time and allowed us to have the furniture installers help setup phones if we wanted too. Staging the phones: pre-configuring them, having the boxes labeled and sorted on the pallet was well worth doing. The wireless phones we signed out to the employees with some other stuff like work shirts. Having the right vendor to walk us through the process was critical.

2. Pilot your install before you deploy it. The environment I was choosing Asterisk for is an automall. Phones are a big part of the business (as with many) and setting expectations is important. We formed a phone users group to have them decide how we wanted to route calls (dial plan), the idea was to get them involved because it is really theirs to use. Some departments were easy and some were not. Sales was essentially create a call groups for the differnt brands we sell and have the operators transfer them to the appropriate group. Service was much more complicated, but having live operators helps a ton. Parts was easy as well, but all of that needs some serious consideration. Knowing you will get it wrong and tweaking it on the fly will happen, do it and move on.

3. We picked Polycom phones and that turned out to be a great choice, the 601's have six "programmable" buttons and great sound quality (handset and speakerphone). The Polycoms have a two port switch built-in and will trunk with the network switch which means the second port on the phone can be a differnt vlan than the phone. So we have them plugged in/wired like this: [network-switch]---[phone]---[computer]. The phones run Cisco CDP, when the switch detects the phone (via CDP) it assigns the phone as a trunk device and allows you to choose what vlan the phone will be on and what vlan the computer port on the phone will be on. Also you can have a differnet vlan if you were to plug the PC directly into the switch. The setup works well and I could go on and on about QoS, edge marking of traffic and PoE issues but I will stop.

4. The FOP (Flash Operator Panel) is a cool thing, but we had to do some customizing for our needs. We looked at Fonalitys HUD, but FOP works great. You can see which phones are ringing, have voice mail (whether it is new or old), transfer calls by drag and drop, monitor the inbound queues and really not have to touch the phone to work the system as an operator. Nicholas, the guy that wrote FOP is an invaluable resource. He was willing to help and has done a great job. I am asking our vendor and am going to make sure he gets paid in some way.

5. Wireless WiFi phones (OUCH): We chose the Hitachi IPC-5000 and Meru Networks for the AP's. Okay I was getting a little cutting edge here, but hey why not?! Lessons:

Meru Networks ROCKS!! They figured out the roaming WiFi thing for sure!

Hitachi IPC-5000's to be determined: it look like either the phones have a high failure rate or we have a bad batch or something. Also it looks like they aren't nearly as durable as say a cell phone/mobile phone (which is VER

The circle is now complete

I am amused at stories like this because this is an example of corporate maneuvers coming back to bite them on the ass. Lots of small companies have been put out of business because software companies have given away products and services in an effort to get market share. Now, the open-source/freeware movement is doing the same thing to the corporations.

Open the Phones

Getting off Cisco CM is just the first step into freedom.

The Cisco 7970G uses XML for its configs and customizable GUI (and HUI) connected to selectable features. Its startup screen has the Java logo. What OS is it running? How do I get it to download and run Java applets? How can I code, install and run native apps?

These little touchscreen phones should offer complete portable offices that even a PHB can use anywhere, without having to search for the "any key". Now that the server is open, how do we open the clients that run on the local HW?

CapEX vs OpEx

Although initial licensing fees, cost of call manager software may be significant (CapEx or Capital Expenses), any large project boils down the OpEx (operational expenses) as these are the costs that can truly make or break a project. What does it cost to support a project over a long term?

Although many of these arguments have been stated against OSS for a long time they still apply here. Technically Asterisk may be just as good as Cisco, but there's an old addage for support "Having one neck to choke." What does this mean? If you have a Cisco LAN, IP core, WAN and telephony, if something breaks you go after ONE vendor. With the OSS telephony solution if the telephony solution breaks, either a) you take on the task of fixing it b) you try to get all your vendors to work together to fix it.

Lastly, professional services can play a large part in a corporations decision to go with a non-Cisco solution. Most large companies probably pay cisco to provide consultative services on their LAN/WAN/Content/optical and so being able to ask cisco "Is my current LAN infrastructure going to be impacted by this telephony change? Or how will VOIP be impacted by this MPLS modification?"

(note: I'm all for OSS, but for a business critical function like VOIP, I don't want to be dependent on the one or two asterisk guys on my staff not quitting. And if the system goes down, I want to be able to have cisco fly an army of engineers on site.

One way Microsoft screwed the tech world

Back in the day, well before software was as complex and complicated as it is today, software was just stuff that came with the computers that were sold. Software itself wasn't the "product." But since Microsoft decided to expand their market beyond hardware makers to consumers and wrote that letter about software piracy, the world changed.

How is this relevant? Again, software is the product. In this case, Cisco and its licensing fees. Most people think of Cisco as a hardware product. While I know it's just a computer with software code that routes information around, it's still, in the minds of many, a hardware product that serves its purposes. But when you are talking about "license fees" you start to think of it differently... more like software. Cisco screwed itself, I think, by moving away from its perception as a reliable hardware product maker. Now you buy their hardware and license the software. It makes people want to shop around more and since the Asterisk product is OSS, well the choice starts to become one of how much money to spend.

It's unfortunate, but seems to be a potentially strong indication of what OSS is doing and why there is such resistance to it, where it comes from and what forms it takes. Looking at it from this perspective shows a nice angle to why software patents are such an important weapon in the software product world.

CCM vs. Asterix

One of the big boons about Asterisk is that you don't have to go through the (overly complex) Cisco ordering process. You get a few Dell servers, smack on Asterisk and away you go!

But... Since the dot-com crash, using the IT guys in any major (non-university) environment to make descisions has been a big no-no. Getting them to even consider a linux-based 'volunteer' PBX for something as critical as telephony when hundreds/thousands/millions of dollars depends on their voice comms is just not reality in my opinion. Call Manager has the Cisco brand behind it (no one is seriously going to question their brand) alongside services such as Cisco TAC and a single-point of contact to bitch and moan. Any financial/enterprise institution can see the benefit of these things and the massive disadvantage in effectively running their own PBX in-house.

While Asterisk is a great product and very interesting to play with (I'm running an Asterisk VoIP solution here), I think the pipe-dream of it cheerfully doing away all the proprietary PBX's is far from reality. Ask any bank.

It's not outside the realm of possibility but...

Don't hold your breath.

I used to do quite a bit of stuff with Asterisk and I happen to run 7 production machines across a DS3 circuit each responsible for up to 4 T1 worth of traffic (92 calls in our current configuration of PRI with 1 DCHAN per span) I am somewhat skeptical about what the boxes would do if they really ran all 92 channels at once It think we have never had more than 2 of the T1's full on a given machine so it's a good thing we load balance them.

I decided I was not happy with this situation so I began work early this year on my own open source soft switch called
FreeSWITCH http://www.freeswitch.org/ [freeswitch.org]

Asterisk is good.

Asterisk is a realy nice package, but not very stable, although it gets better with each release. As in, it "works", until suddenly... you get dropped calls. Or the incoming caller gets a high-pitched squeal into the phone instead of your PBX menu.

Of course, a lot of these issues have more to do with the zaptel drivers, rather than Asterisk itself. But trust me - you WANT to stay up to date with the Asterisk releases. Do not run anything below 1.2.X.

Re:They went just a little bit too cheap....

I'd be interested to see the total cost of ownership including ongoing maintenance of Asterisk vs Cisco which has an abundance of specialists.

Re:Easy there, Cowboy

"Know" for "no" is a truly boneheaded error, much more so than "then" for "than".

You forgot a question mark in your post correcting him - that's an even more bonehead error.

His message was adequately communicated - you don't need to be annoying and correct him. If you were adding clarity to his post, it would be one thing, but you are just nit-picking. Add something to the conversation or go the hell away.

Re:They went just a little bit too cheap....

What the hell is wrong with Dell servers? They are the #1 vendor, for a good reason. Who exactly makes better servers?

Re:They went just a little bit too cheap....

Doesn't surprise me coming from an academic background. Business in the last few decades has had a lot of gall saying colleges have to learn to be lean like they are. I know in our state money to the state colleges has been stagnant for years after a year of significant decline. You might think your tuition pays to keep a college running but it doesn't. Working at a couple private colleges I have become acutely aware that they are charities and walking through a private college campus in particular is like walking through a mausoleum. Every building, every rock garden or artwork, heck, probably every TREE is in memory of some dead person. It isn't surprising when you see even a state-subsidized college or university taking the initiative on something like this because their intellectual talents _are_ their most fluid assets. And making due with the hardware they can afford. If it bothers you and you are a Sam Houston grad, I suggest making a dedicated donation to your alumni association.