Welcome to VoIP! The importance of the role an SBC plays in a VOIP topology is somewhat based on what type of VoIP service you're offering. If you're basically providing a last-mile POTS service from a CO in your region of interest and then trunking to a VoIP provider for off-net access you could somewhat justify against purchasing an SBC and setting up properly configuring access lists on your IP edge (you should also ensure whatever core/edge Ethernet/IP routers/switches are capable of handling the throughput you require). However if you're providing phone service over the public internet using IADs/ATA/IP Phones then you will absolutely want to look into SBC options as they provide various important functions such as NAT traversal and load shedding/DOS protection which is important when opening any type of interface to the public internet. Whether you go with commercial hardware/software solution or go with open source should be based on cost and support. If you are unfamiliar with VoIP it may be a better path of approach to go with commercial vendors since they can provide you with support where required rather than having to dig through various mailing lists and debugging exercises to narrow down what may or may not be a bug in the end. In contrast, if you want to implement new functionality, with some talented programmers you can do so yourself with open source solutions vs. going through long and drawn out battles/cordial discussions with vendors to get them to implement some functionality you may find useful or novel. Regards, Justin Randall Team Leader - VoIP Engineering Comwave Telecom Inc. From: voiceops-bounces at voiceops.org [mailto:voiceops-bounces at voiceops.org] On Behalf Of Jonathan Jones Sent: June-21-10 1:56 PM To: Matthew S. Crocker Cc: voiceops at voiceops.org Subject: Re: [VoiceOps] New to VoIP On Mon, Jun 21, 2010 at 1:38 PM, Matthew S. Crocker <matthew at corp.crocker.com> wrote: Everything is software, a 'Session Border Controller' is software. All software needs some hardware to run on. You can get a SBC that runs on Linux/PC server based hardware or you can get a SBC that runs on dedicated hardware. Your softswitch is going to need several servers to run on as well. The smallest Broasoft/Broadworks implementation is 2 servers for the softswitch running virtual servers and 2 SBC (Acme Packet). You can't get into that type of a setup anywhere near $50,000 though. For 2000 subscribers I would highly recommend you partner with an existing VoIP provider and resell their service. They can handle all of the VoIP heavy lifting and send you the Call Detail Records which you can then use to bill. If you want to do it yourself you'll probably have to live in the open source world to keep below the $50,000 limit. Thanks, I had been looking at the softswitch software route and had been quoted a price of $12,288.00 for 1024 simultaneous calls. So I figure $20k more or less for a softswitch that would do 2000 concurrencies, another $10k in hardware. Am I too naive to want a software based solution to perform least cost routing, billing, perhaps calling cards, etc. I've been looking into either Asterisk or Callweaver, OpenSIPS and Soft-Switch.org for signaling. Am I again, naive or can it be done. If not I don't want to waste my time nor money especially if I can do it the right way.

Peter R. wrote:

Don't take this the wrong way but VoIP isn't best effort data. There's WAY more to it than a couple of open source devices. There's billing, E-911, directory assistance, LNP, DID's, CNAM, SS7, routing and trunking, and Insurance. There are folks on this list that still have frequent issues. Partner with someone that knows telecom and knows VoIP. It can't be done well on the cheap.

This is the best reply yet on this topic. -- Carlos Alvarez TelEvolve 602-889-3003

On 6/21/10 1:38 PM, Matthew S. Crocker wrote:

Everything is software, a 'Session Border Controller' is software. Well mostly. As an example, Acme Packet SDs have very expensive (comparatively speaking, at least) content addressable memory (CAM) wherein lives the lookup table for all active SIP sessions, thus allowing the hardware to deal with a much higher volume of RTP packets than a general purpose computer of the same general size and "power" could hope to process, not to mention making it more resilient to DoS attacks.

Just like with, say, routers, you can do wonderful things with a general purpose computer until you hit a capacity barrier, where you realize that there's yet one more reason that people send Cisco boatloads of money for high-end routers. ;-)

For 2000 subscribers I would highly recommend you partner with an existing VoIP provider and resell their service. They can handle all of the VoIP heavy lifting and send you the Call Detail Records which you can then use to bill.

He said 2,000 concurrent sessions, which is rather larger unless is town is occupied solely by call centers squeezing out every penny from their carrier. Actually, that's a suspiciously big number--maybe he meant 2,000 subscribers. I'd worry about who is going to do the customer service, not to mention what level you want to live up to, early in this process. That's likely to have more to do with customer satisfaction and consequent retention than whether you outsource your voicemail server.

If you want to do it yourself you'll probably have to live in the open source world to keep below the $50,000 limit.

If you go fully commercial, high-end, with some redundancy, you can easily blow through a million before you're ready to turn on the first customer. I quite agree with Matthew on your two basic choices. To the list of open source switches you're looking at, I'd add Freeswitch. And, yes, it can certainly be done with some of the products you've mentioned (I'm not familiar enough with some of them to have an opinion). It doesn't come with a nice SE to help you design everything at "no extra charge" like Broadworks would, but it can certainly be done. I know of one small provider who uses Asterisk and a bunch of their own software, with a staff of 3ish, one of them spending most of his time as a programmer doing his best to provide the customers with a nice experience in their custom, web-based front end. --Jon Radel

From my experience setting up a CLEC using open source software is not the way to go, and I am typically all for open source and design much of my ISP side around it. If you actually plan on growing I would recommend a true switch though, it simply takes too much time to roll your own solution and it will never be as reliable as some of the gear out there. However, you will also end up spending a minimum of $300k I would estimate.

-Eric

On 06/21/2010 02:49 PM, Jon Radel wrote:

...

Well mostly. As an example, Acme Packet SDs have very expensive (comparatively speaking, at least) content addressable memory (CAM) wherein lives the lookup table for all active SIP sessions, thus allowing the hardware to deal with a much higher volume of RTP packets than a general purpose computer of the same general size and "power" could hope to process, not to mention making it more resilient to DoS attacks.

CAM-like memory, and also ASICs to assist in processing of media and the CPU work involved in packet forwarding.

There's no question that pound-for-pound, a high-end SBC can handle a whole lot more concurrent calls (with media) than general PC hardware with a general purpose task-switching OS with a general purpose PCI bus, general purpose I/O scheduler, and a userspace process doing a lot of the work.

However, a) you may not need that and b) depending on what it is exactly that you intend to do, it may be quite possible to design a topology that does not involve relaying media at all most of the time (which is where proxies tend to be useful), in which case the case for the open-source option is substantially bettered.

-- Alex Balashov - Principal Evariste Systems LLC 1170 Peachtree Street 12th Floor, Suite 1200 Atlanta, GA 30309 Tel: +1-678-954-0670 Fax: +1-404-961-1892 Web: http://www.evaristesys.com/ _______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops

In the fullness of time, we have discovered the billing is a big, thorny, gnarly deal to do correctly. Taxes, surcharges, FUSF reporting and remittance, FCC forms. Just to do it correctly chews up head count and time. Would recommend using an outside voip billing service if you can at the start. -----Original Message----- From: voiceops-bounces at voiceops.org [mailto:voiceops-bounces at voiceops.org] On Behalf Of Eric Hiller Sent: Monday, June 21, 2010 2:25 PM To: voiceops at voiceops.org Subject: Re: [VoiceOps] New to VoIP

From my experience setting up a CLEC using open source software is not the way to go, and I am typically all for open source and design much of my ISP side around it. If you actually plan on growing I would recommend a true switch though, it simply takes too much time to roll your own solution and it will never be as reliable as some of the gear out there. However, you will also end up spending a minimum of $300k I would estimate.

-Eric

On 06/21/2010 02:49 PM, Jon Radel wrote:

...

Well mostly. As an example, Acme Packet SDs have very expensive (comparatively speaking, at least) content addressable memory (CAM) wherein lives the lookup table for all active SIP sessions, thus allowing the hardware to deal with a much higher volume of RTP packets than a general purpose computer of the same general size and "power" could hope to process, not to mention making it more resilient to DoS attacks.

CAM-like memory, and also ASICs to assist in processing of media and the CPU work involved in packet forwarding.

There's no question that pound-for-pound, a high-end SBC can handle a whole lot more concurrent calls (with media) than general PC hardware with a general purpose task-switching OS with a general purpose PCI bus, general purpose I/O scheduler, and a userspace process doing a lot of the work.

However, a) you may not need that and b) depending on what it is exactly that you intend to do, it may be quite possible to design a topology that does not involve relaying media at all most of the time (which is where proxies tend to be useful), in which case the case for the open-source option is substantially bettered.

-- Alex Balashov - Principal Evariste Systems LLC 1170 Peachtree Street 12th Floor, Suite 1200 Atlanta, GA 30309 Tel: +1-678-954-0670 Fax: +1-404-961-1892 Web: http://www.evaristesys.com/ _______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops

_______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops This message and any attachments are intended only for the use of the addressee and may contain information that is privileged and confidential. If the reader of the message is not the intended recipient or an authorized representative of the intended recipient, you are hereby notified that any dissemination of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail and delete the message and any attachments from your system.