[VoiceOps] Which Softswitch?

[Alex Balashov]

Indeed, it doesn't seem to me that open-source systems are the thing to 
be avoided, nor that it's necessarily possible to do so. Moreover, the 
value proposition and trade-offs of open-source systems are quite clear. 
It seems to me the largest long-term value is in integration paths and 
connectors; most proprietary, "big iron" boxes just do what they do, and 
that's all they do, more or less. They may have a lot of features, but 
that's the feature set, and tying it together into novel, innovative and 
commercially differentiated third-party services is hard.

That said, I think we all know the sort of open source-based system to 
which the OP was referring. Asterisk and FreeSWITCH are low-hanging 
fruit, and have invited a lot of bad implementations and poor 
architectures. There's nothing wrong with using these systems 
foundationally within a carrier-grade product, as long as the system is 
architected correctly, in a horizontally scalable, distributed and 
fault-tolerant way, and that's a fairly complex undertaking of software 
engineering.

Vendors of these kinds of solutions also often do not provide a level of 
support that comports with telco sensibilities; their reasoning is 
either that the customer should largely support it themselves, since 
it's all built on open-source components, or their scope of support is 
narrow. Consistency and commitment can be an issue.

I can only speak firsthand, but in our case it has been very clear to me 
since the early life of our open source-based, commercial ITSP product 
that customers expect a high level of service value, and that the vendor 
relationship, along with the institutional domain knowledge and 
expertise provided, is as much a part of the value proposition as 
software itself. It's also been very clear that they expect support for 
the _entire_ technology stack of which the product consists, much as 
they would receive from Acme Packet or Sonus. Our customers don't care 
that our product ties together Kamailio, SEMS, PostgreSQL, Node.js, 
Redis and, ultimately, Linux, nor do they care about the degree to which 
we can or cannot exert direct control over bugs in these third-party GPL 
components. They expect us to configure the installations, maintain 
them, and troubleshoot, debug and fix as necessary.

I don't think this insight is necessarily common among vendors of open 
source-founded products. I've heard a lot of things like, "Oh, well, 
that's a bug in Asterisk, that's not a problem with our application." If 
the vendor sells and supports an Asterisk-based platform, to a large 
extent, it should be the vendor's problem. They may not be able to 
resolve it themselves, but they should own it, communicate it 
efficiently to the appropriate parties through expedient channels, and 
marshal the appropriate resources in support of fixing it. Not 
everything is always possible, of course, but many things should be 
possible most of the time.

-- Alex

-- 
Alex Balashov | Principal | Evariste Systems LLC
303 Perimeter Center North, Suite 300
Atlanta, GA 30346
United States

Tel: +1-800-250-5920 (toll-free) / +1-678-954-0671 (direct)
Web: http://www.evaristesys.com/, http://www.csrpswitch.com/