[quote]Metropolitan Health Group (MHG) had saved R4.5 million (~500K Euro/650K USD) in the last two years by switching to open source software for its Government Employees Medical Scheme (GEMS) call centre. This is according to MHG IT infrastructure manager Rod Russell, who said they migrated to Asterisk when they had to deliver a fully functional call centre in just two weeks.
Working with Connection Telecom, the MHG IT team created a 40-seat call centre in the required time.
MHG administers 21 medical schemes across the country and has 130 out of their 570 seats running on Asterisk. Russell said they are looking to migrate more seats to Asterisk over time.
Starting with the basic Asterisk solution, Russell’s team worked with Connection Telecom to develop additional functionality specifically for the call centre environment such as call waiting facilities, an inbound, outbound call tracking and other operational data facilities.[/quote]
No. But one area where Digium needs to assert more effort is collecting and disseminating case studies such as these to the community and market at large. Spread this type of news far and wide…
Hi MM,
Asterisk would be adapted more if it uses the Softswitch design where
the transcoding part is seperate from the call handling part.
Such a design is scalable and reliable. Also an SNMP mechanism can be
easily adapted in such a design.
What do you think?
[quote=“vinod.vijayan”]Hi MM,
Asterisk would be adapted more if it uses the Softswitch design where
the transcoding part is seperate from the call handling part.
Such a design is scalable and reliable. Also an SNMP mechanism can be
easily adapted in such a design.
What do you think?[/quote]
Neither SIPX nor Freeswitch have the size of the Asterisk community, therefore it is a question of momentum. Although SIPX went in as the PBX at Amazon for ~5K stations…
I am not denying the popularity of asterisk.
What I meant is that there is room of improvement regarding
Seperation of call handling and codec translation
Codec translation is done by MG in softswitch while MGC handles the call processing. Because of this this both parts can be seperated. There can
be a distributed solution in which there would be many MG’s under the control of an MGC. Also MGC to MGC communication for improving scalability.
The communication between MG and MGC can be by means of a IMS recommended protocol like Megaco.
What this can achieve in Asterisk?
The codec translation can be distributed among different systems such that the dsp processing is distributed. Asterisk can handle as many calls as possible.
An SNMP application can be built that will give information about the resources in use. This can help in finding the threshold or finding which
calls(conferences mostly) consume most resources.
Calls can have priority such that important calls in case of overload can continue.
Well, these things take time. Look at the linux kernel. This is a 150 year old industry. And from my experience, you don’t nail a perfect design first go - iterative process. I wrote an implementation for a CSTA stack - 3 times before I got it right - and by right I mean effing great!!! So much more maintainable and stable. And the first two times were in greater production use than my own effort - but that’s business brains as opposed to techy brains - same deal for case studies for asterisk.
Digium will get around to it eventually, they’re not stupid. The limitations create work for a bunch of people. Just like the GUI interface - great for the simple solutions, but not so great (yet) for your ridiculous call flows that some customers ask for.
The sip channel is being overhauled by some dude called olle and in 12-18 months it will kick some axx more than the current implementation.
And for the next 12 months we’ll all have to put up and bear with all the limitations of the open source community too - we don’t have the luxury of sitting a few metres away from each other and getting our heads working.
So, get into the developer channels with your expertise and push a few diagrams on some website and try and bandy a group of uni students or open source junkies that want to prove their worth to implement an improved design. Seems like you know what you’re talking about - give us a diagram to explain to the not-so-interested-in-the-architecture so we have something to print off and pass on or email to someone with the skills, time, know-how and “circle of friends” to get some stuff happening.
Dia is a good little program for drawing things up, so is freemind off sourceforge.
Look forward to a reply - a picture is worth a thousand words
Peace
Besides, with an OpenSER/Asterisk combo you may achieve much of the separation required for ITSP type scale. Although, many purport to have that with Asterisk alone anyway…
I am currently in the process of project managing an implementation of Ericsson Solidus eCare for a 60 user call centre. The customer already has an Ericsson MD-110 system but the implementation (inc. On screen CCA software, IVR, and on screen wall board software) is going to cost between ?75K and ?100K.
Asterisk has huge potential in this market but needs credibility. If only there were some reference sites, especially in the UK, who use Asterisk for call centre system this would help to show customers that it really is a robust solution.
I have worked with Asterisk for a couple of years now and find it excellent but I need to convince my customers and they won’t just take my word for it. The best way is to get them to speak to someone else in the UK who has already done it!