So the objective seems to be to make the “system” resilient even though the individual components might not be. I suppose this is how successful human civilization works, individuals might die off but as long the culture is resilient it survives and even prospers.

Regards,
Mick

Great example of the industry behavior. There is nothing wrong with “Carrier Class” systems and protocols or their expectations but it always seemed strange to me that many believe that to have this type of high availability and scale, you must build and implement technology exactly as it had been implemented in the past where that was the outcome. To hold this believe is to think that no future technology can be as or more resilient or scalable as the past…a pretty silly belief as it is betting against innovation and the future.

I remember having a dialog with some folks who built 6 9′s hardware and we got on the subject of resiliency architectures. My observations is that even at 6 9′s, the MTBF is still not infinity so every system they build will eventually fail. If we acknowledge that then the innovation shifts from simply making the elements resilient to focusing on system resiliency and behavior when the failures occur. When you think of the system then suddenly you can trade off a lot of hardware and node complexity for more effective inter-node protocols and distribution of behavior that survives in the presence of a failed element. I think we are seeing some of this in the current hype on cloud computing where people were attracted by the economics and flexibility but now are beginning to focus on innovation to make the “cloud” resilient even though the nodes are not.

There is nothing wrong with the past and its technology but an open mind goes a long way in creating a better future technology set.

John

Today we were testing some call flows originating from a IP PBX, through a session border controller onto a softswitch, connected to the PSTN and SS7 network. My co-worker and I work on the IP side and our testing partners work in the pstn and SS7 realm.

There was a particular test case where the SS7 tester was walking us through a large bunch of SS7 parameters (way beyond me). My co-worker remarked that it was so complicated that it would be nice to migrate fully to SIP.

The SS7 tester rebuked him, saying that SIP was incapable of addressing all the intricacies of our network and only when it starts incorporating the SS7 parameters could we ever consider using it in a production environment.

As I sat there taking log traces on the SBC listening to this, I wondered about the VoIP network I helped build back in 2004 that operated for 2 years before our newly hired executives killed the service for it not being built on carrier-class equipment. Oh yes, these executives were ex- Telco types (from the another large Canadian telco) they were hired to portray a veneer of telco goodness for potential buyers.

Oh well, I’m sure some day we will get a next gen voice network here in Canada.