What is Digg?85 Comments
- frogman54, on 04/09/2008, -0/+28This is the kind of thing where you get all excited and you go to tell all your friends, "Hey, I got to play with an open source class 5 telephony switch today!" And then they all give you that "you are such a nerd" stare.
- CrushThemTorg, on 04/09/2008, -0/+21A terrifying new form of fanboy emerges on Digg. Thusly the PBX Wars begin.
- a1lostnomad, on 04/09/2008, -0/+20Congratulations on RC1! A couple programmers in my office are developing a VOIP solution with FreeSwitch and it is really amazing. In just a couple weeks they've setup a complete phone system (with features most offices only dream about), and integrated nearly every aspect of the system into our web based software. Open source FTW!
- FKnight, on 04/09/2008, -2/+15Yes, that would be so awesome. Everyone with an Asterisk or FreeSwitch box in their house ...... and no cables, wires, optical fiber, or antennas for them to talk to each other over.
- Damin, on 04/09/2008, -1/+13Congrats!
- jdfmcok, on 04/09/2008, -0/+11I've used it heavily and it rocks compared to Asterisk. I've got it running 12 phones in my office right now. The learning curve from Asterisk to FreeSwitch wasn't too bad, and freeswitch runs great on osx, windows, and linux.. where Asterisk only ran on linux.
- anthm, on 04/09/2008, -0/+8That's a perfectly reasonable point of view. Being the author of the software this article is about, I can assure you, we have no intent to make anyone use it. We are just making it available for anyone who is interested.
- cronparser, on 04/09/2008, -1/+8awesome
- branchcut, on 04/09/2008, -0/+6also you can write IVR's in python and control it over sockets with the python client library
- serif69, on 04/09/2008, -2/+8Waiting for the forum post with complaints about issues with the drivers, followed by a 200-post bitch session, followed by a single programmer patching it, followed by 20 other programmers ***** it up trying to migrate the patch to other telephony switch distros. Lather, rinse, repeat.
- inactive, on 04/09/2008, -1/+7Answer: When companies have an existing, paid for solution to a vital service, they don't swap software and servers on a whim.
- SwK386, on 04/09/2008, -0/+6yeah what he said! this is completely True... The Switch is a major component yes, but you arent going to run out tomorow and buy a DMS100 and call yourself a carrier either... you're going to spend another bus load of cash on integration, billing, provisioning, account, more engineering, more integration, etc
But removing that one $500K line out sure makes it easier! - lanwifie, on 04/09/2008, -0/+6The guys that create and maintain freeswitch are not just geeks wasting time ... they are very serious about making an opensource product that will revolutionize the industry.... Support??? All the support needed to install, run, maintain FS can be found in the FS irc channel. Not only do these guys create a magnificent switch that out performs most others in production situations but they back it up by helping people when they come with questions.
Reliability, Great Support and Documentation AND IT's FREE ....
What more do ya need?!?!
~Lanwifie - lohphat, on 04/09/2008, -2/+8I could drain a cocktail bar telling you how many times geeks (which I am one, just grayer) try to tell me how to run infrastructure when they have ZERO experience. They cannot (or will not) learn the difference between "local optimization" (to fit their specific skills and comfort) to the more economical "global optimization" (i.e. economical and can be used across a greater spectrum of skill set found in the real business world where not everyone's a geek). There is great value in "it just works" and when it breaks (not if) we can call someone with a 4 hour response who knows how to fix it. There's no value in paying full-time salaries for in-house personnel to remain experts in systems which rarely break. It's like paying a dentist, doctor, cop and fireman to live with you in case they're needed. This is completely lost on people who are not the one whose job is on the line when things fail.
- HonoredMule, on 04/09/2008, -0/+5"If you have to develop the subject matter expertise in-house to support OSS projects to the level of commoditized out-sourced services then you've not worked out the numbers; you're then usually spending money on in-house skill sets than can usually be had for a fraction from a vendor."
Well then it's a good thing there are many vendors providing complete packages (hardware, installation and setup, additional features like a Do Not Call blocker and predictive dialer, support contract, etc.) around OSS like Asterisk at a fraction of the cost of just getting the hardware for traditional proprietary systems.
In case you hadn't guessed, I work for such a vendor...and downtime is virtually non-existent. Our full-time support staff does sysadmin stuff and browses slashdot while waiting for support calls (an endangered species here). Our weekend support is a complete waste of our clients' money (although they obviously value that availability) because without staff actually using the systems, there aren't even occasional minor glitches to call about. - jmhunter83, on 04/09/2008, -0/+5man brian, from just doing an implementation of *odbc to writing a whole system
+2 if i could - FKnight, on 04/09/2008, -3/+8Wow. You got Dugg down for knowing what you're talking about. Welcome to Digg. Nevermind the fact that you obviously have experience, knowledge, and expertise in business management. A bunch of computer geeks on Digg who hate software companies know what's best for a business.
- aksyn, on 04/09/2008, -0/+5Sorry, this pushed a button. Actually plenty of business critical needs, my own included, are being met by FreeSWITCH every day. We've been using using it in production since well before RC1, and I would bet my job on its reliability.
Anyone who's used Asterisk in production knows that it has some serious stability issues with >200 channels, screwing whatever savings you thought you were going to get by "going FOSS" by having to buy racks full of servers and people to baby-sit them. Version numbers mean nothing at the end of the day, but I think that it's /taken/ so long to get to even RC1 should give you an idea of how crazy about perfection Anthony and the team are.
We pulled Asterisk out and replaced it with FreeSWITCH just before launching our initial beta. We didn't do it on a whim, we did it because Asterisk just wasn't up to scratch - and if you take a look at the FS source, play with it in the lab and stress test it right alongside against Asterisk, it makes the latter look like a cheap toy. It scares me that I know of at least one major carrier using Asterisk in the background for a 999 service (another meaning to the term "deadlock"), when I've seen it freeze playing music-on-hold to just a couple of hundred channels in a lab.
Yes, it doesn't have the market share that Asterisk has yet, but that's really just a function of marketing prowess (something I'm sure will come with time) and not a testament as to the whether a product is ready to compete.
Commercial support contracts (and I mean real support) are coming, and very soon from what I hear. Way before it even reached RC1, the team were flying out to people who are running this in production at far greater levels than we are, and making sure it gets the job done. I've seen it running alongside TelcoBridges TDMVoIP boxes, at price points that would historically have been laughed out of the room in the TDM world. If I were the "big vendors" I'd be pricing up those solutions and wondering how I'm going to get new customers and send my kids to school next year.
Projects like FreeSWITCH are shining examples of where dedicated people, and open source projects are revolutionising entire industries - this isn't just for small businesses. From what I've seen (and I've been in this industry a while), FreeSWITCH can already hold it's own against the big vendors.. imagine what it'll do when it reaches version 1. - mercutioviz, on 04/09/2008, -0/+4Great project, great community!
- irshadigg, on 04/09/2008, -0/+4Great Job... Great people... Great outcome!!!!
- tmissner, on 04/09/2008, -0/+4Built a load test platform to drive an IVR app in a day and half. Replaced similar system running on 7 boxes built with sipp ( streaming rtp ) and asterisk ( answering B leg call ) with 2 using FS. Scalability is unparalleled.
- linagee, on 04/09/2008, -0/+4Why not hire a support vendor to support your OSS switch?
- inactive, on 04/09/2008, -0/+4"In theory, FreeSwitch could become the platform for a new carrier, but there’s far more to building a carrier than just purchasing a switch. Purchasing and integrating the proper billing and management systems are bigger changes."
Uh huh... - vroom101, on 04/09/2008, -1/+5Excellent news for an excellent project!
- mochouinard, on 04/09/2008, -0/+4Good job, Near Version 1 it GREAT news
- lohphat, on 04/09/2008, -0/+4Yes because that''s the most economical. Otherwise having to staff experts sitting around mostly doing nothing except during the occasional fire-drill is $$$. Again, this is the same reason we outsource dentistry and surgery, there's no point is staffing your house with these skills -- flossing and aspirin (your analogy to the bell ringing monkeys) is OK, then you call in the experts when you need them.
- YodaJones, on 04/09/2008, -1/+5Death to telcos. Imagine if everyone ran this and Asterisk. Imagine a world without telcos.
- elementop, on 04/09/2008, -0/+4Saying that it is always more cost-efficient to outsource than to DIY is just as fallacious as saying it always more cost efficient to DIY rather than outsource. Where I work, we use a lot of open source software because in our case, it really *is* more cost efficient to DIY. But in those cases where we have found vendors who can do it better/faster/cheaper than we can, guess which option we chose?
Case in point: we have a network built with Cisco routers -- really expensive, really inflexible, really buggy, and oh, did I mention they were really expensive? There was a bug in the software that crashed the network within one week of our installation. We were on the phone with Cisco for *eleven hours* after that first crash. It then took something like NINE MONTHS to get a patch for this particular bug. OTOH, we have a vendor (ImageStream Solutions) that builds routers based on Linux using Quagga or Gated (Gated sucks, but Quagga is really, really good). After over three years in service, we just had our *first* hardware failure. And, although we've found some bugs in the software from time to time, ImageStream usually has a patch available within 24 hours of us reporting the bug. I think once we had to wait a whole weekend, but even that is orders of magnitude better than the 9 month response time we got from Cisco.
Our business uses a lot of open source software, some supported by vendors, some not, and we wouldn't do it any other way because it meets our needs more efficiently than the alternative. - StinkBait, on 04/09/2008, -1/+5So, what you're saying is you want a bunch of "ring the bell monkey" types to watch the network, and you call in the "real engineers" when it breaks.
Sounds like every place I've ever worked. - shoeman22, on 04/10/2008, -0/+4I hear your argument related to implementation costs and if your company is already competent w/ a given system and/or you already have the hardware, then it probably doesn't make sense to throw your system into an upheaval for any new system unless it is really necessary.
That being said, I struggle with the TCO argument overall, and it certainly doesn't apply for my old company. I recently worked for a small VoIP company and we used a proprietary $100k+ system to handle our calls. In addition to the initial charge, it costs us roughly $25,000 in maintenance agreements and updates and to top it off, we still had to pay $35/license. To contrast, we recently started hosting some customers on a new Asterisk cluster, $5,000 in hardware and $0 in licensing fees. According to our network engineer, the Asterisk cluster has actually been slightly more reliable than the proprietary system! How can you beat that? For well under the $95,000 difference on the initial outlay, plus whatever licenses would have been, and maintenance contract fees, we could definitely pay to train our engineer to become an expert in Asterisk as well.
Another angle to consider as well, especially for companies in the telecom game, is if you're using uber proprietary systems like we were, it is very difficult to branch out to different markets due to very high initial investment just to test a market. If you're using Asterisk, or FreeSwitch, you can probably rent space in a datacenter and test the market for around $10,000 - 15k which is very reasonable. We actually attempted to branch out and went with a different move by buying thick pipes to transfer calls from the location to to our home server, which saved us from a 100k outlay, but did saddle us with a nice bandwidth bill every month that would have been absolutely unnecessary if we had installed an Asterisk cluster at the location. - anthm, on 04/09/2008, -0/+4you can find that at http://bugs.digium.com
- briankwest, on 04/09/2008, -1/+5That depends on if the solution is actually working or not!
- linagee, on 04/09/2008, -1/+5antennas need telcos? mesh networking ftw.
- bryxal, on 04/09/2008, -1/+5Free as in speech...
Think of it as Free Speech. "Liberty of software" - elementop, on 04/09/2008, -0/+4True, but when it's time to upgrade/replace, it might be a good idea to consider alternatives to what you've been using.
- aksyn, on 04/09/2008, -0/+4Well, let me clarify something. I bashed Asterisk earlier, but believe that I'm not alone in my grievances (certainly the Digium bug tracker would seem to agree with me). Asterisk was however instrumental in us being able to develop our product in a very short space of time (at the time, FreeSWITCH simply wasn't on my radar and Asterisk was the de-facto way you did such things inexpensively).
It's only when I came to bring Asterisk into production, with load testing and whatnot that its warts started to show.
I'm sure Anthony could tell you the story more accurately, but I believe Asterisk was originally conceived to replace those expensive Avaya, Nortel PBXs with commodity PCs. This, it does extremely well, for relatively small values of n (where n is ~250 concurrent calls per server).Bolt-ons like OpenSER let you scale Asterisk up past the point where it has threading issues, spread it across multiple machines and provide some level of redundancy.. but the core of Asterisk was never really designed to handle massive call volumes, and could never be coaxed into doing so without a major rewrite (which, I believe, is what was ultimately responsible for Anthony "giving birth" (sorry) to FreeSWITCH).
To answer your question.. did we lose a lot of time in going down the wrong route initially? No, fortunately we'd abstracted away all of the logic for call handling into our erlang application, and just had a thin control layer above the switch/softswitch so that the switch could do what it does best. Which is to switch calls.
We were a little upset that Asterisk was a bit poor at this, and I apologise to Mark if I'm being harsh on his product (although I don't think he's doing /too/ badly out of it) but I am extremely happy that we stumbled upon FreeSWITCH.The core team, the community, the expedience of bug fixes (they've been known to actually accept patches!) puts it above the other offerings at this time.. however, there's nothing like a little healthy competition, so maybe Asterisk has a v2 I don't know about in the works :)
Did we waste any money in switching? No, we didn't have vendor lock in - Asterisk was open source, replaceable, malleable.. but it just couldn't cut the mustard. If something else comes out that's better than FreeSWITCH, maybe we'll switch to that - that's one of the beauties of open source (so long as you loosely couple things).
FreeSWITCH gives us the kind of performance we were expecting to get out of a softswitch, at a price that enabled us (and many others, with enormous thanks to Asterisk for also enabling this) to innovate in what has typically been an extortionately expensive industry to break into. "Known VoIP solution" - I'm not sure what you were referring to, but if we're talking about the Huwawei's of this world, then we simply wouldn't be here today. - pnmoore, on 04/09/2008, -1/+4What about every house with these kinds of systems and wireless capabilities that mesh with their neighbors? A redundant voice network with each house being a node in the mesh.
I know that is a pipe dream (still need long haul capabilities to communicate for greater distances than your neighborhood), but it would be cool. - datagog, on 04/09/2008, -1/+4A revolution is a revolution, regardless of where applied. Anthony and team, like a lot of open source proponents, pioneers and refiners are to be applauded. There's always pressure to bring prices up or down, from various gated and weighted interests, but whenever intelligent and articulate people of this caliber set out to out-innovate literally everyone in the market we all win. Whether the impact is felt directly or indirectly, they are driving technology and progress forward. Forcing the entrenched players to also work harder, faster, and smarter, well think of that as a bonus. I work for one of them and I do.
- pnmoore, on 04/09/2008, -1/+4If you really don't know then this doesn't interest you. Move along...
- anthm, on 04/09/2008, -0/+3Thank you
- briankwest, on 04/09/2008, -0/+3Well this is in my test lab.. If you look at SIPx it can manage all kinds of vendors and comes with a GUI. That works great for many people too.
- cybrsrfr, on 04/09/2008, -0/+3I've been looking forward to Freeswitch release for a year and the release is almost here! I'm using Freeswitch RC2 on FreeBSD7 and it is working great. A phone geeks dream come true!
- ggidster, on 04/09/2008, -1/+4Great job guys - nice piece of work!
- pnmoore, on 04/09/2008, -0/+3Your points are all valid, but I don't recall anyone saying this was the end all/be all of replacements for all class 5 switches in the world.
I am sure you also know that there are the right tools for the right job in any situation. It may be that a small business that wouldn't normally be able to even buy the commercialized version could use something like this for some non-mission critical application that allowed them a fall back to normal phone service if a failure occurs. In this case it could be a very low cost solution to solving a minor business need that isn't worth paying for the support and up front costs associated with the big vendors out there. There are tons of business needs that every company has that will never be resolved without a low cost solution being available. The ROI just isn't there for minor issues. For anything business critical you would be crazy to use something like this, especially since it is only RC1. - linagee, on 04/09/2008, -1/+4Depends if it's a new company who does not have any servers yet.
- briankwest, on 04/09/2008, -0/+3I use Snom, Polycom, Aastra and Grandstream so far and it works great with FreeSWITCH.
- briankwest, on 04/09/2008, -0/+3Its not fully supported on those platforms.... I have had it on OS X but it sucks.
- iXneonXi, on 04/09/2008, -0/+3Friends, what friends?
- lohphat, on 04/09/2008, -0/+3What do you do for phones? I know the Polycom ships a SIP phone but what do you use for rank-and-file 1/3/6 line phones? I've used ShoreTel in the past (really easy to maintain and great reseller support) but the downside was you had to use their phones.
- lohphat, on 04/09/2008, -0/+3"We pulled Asterisk out and replaced it with FreeSWITCH just before launching our initial beta."
How would you quantify (in both time and money across everyone affected) spent on Asterisk before changing it all out for Freeswitch vs. having just gone with a known VoIP solution? -
Show 51 - 88 of 88 discussions
© Digg Inc. 2009
— Content created and posted by Digg users is