[asterisk-biz] Looking for carrier grade redundant solution
Christopher Aloi
chris.aloi at gmail.com
Sat Jul 29 10:28:20 MST 2006
Thanks for the informative replies.
While Enswitch sounds like a a nice solution, I don't believe it's for our
shop. I really would like to develop this solution internally.
My thoughts thus far are pointing towards a scalable, redundant solution
based on stock hardware.
I am thinking of running a matched server pair (1 or 2U) as the application
servers (both running a matched Asterisk build) most likely on RHL. The
servers will contain redundant hardware (power supplies etc...)
The application servers will be backed up by a matched pair of MySQL servers
which will use MySQL replication between each other and will use a floating
IP; a heartbeat will beat between the pair and nominate one a master should
another fail.
The disk storage will be local or offloaded to an existing netApp filer.
>From the reading I've done, running SER as a proxy to the two application
servers should serve it's purpose well, providing load balancing and
monitoring of the application server behind it (should one Asterisk box
return a 3XX-5XX I would then route advance to the second application server
and take the first offline.
Question - Is running SER by best bet here? Would I be better off running a
heartbeat between the two app servers?
I am not too worried about SIP registration/expiration's of the SIP users or
peers:
1) Most of the call center traffic will be terminated out another platform;
the calls are being sent to a DID (not a SIP URL).
2) My SIP peers are all within the trusted network core and will not need to
maintain status of each other.
Here's where it gets tricky:
As agents login to a queue or become 'available' Asterisk is maintaining
their state; I believe an agents state is maintained across an Asterisk
reload. Could an agents state be maintained should the second app server
take over for the first?
Anyone attending ClueCon next week??
Thanks,
_Chris_
On 7/28/06, Douglas Garstang < dgarstang at oneeighty.com> wrote:
>
> What about sip registration replication?
> What about SIP subscription replication?
> What about BLF replication?
> What about using DUNDi to replicate applications for redundancy?
> How would you handle different phones ability to failover if they don't do
> it so well?
> How would handle the fact that the config files have a hard coded database
> IP?
>
> And so on...
>
> I don't think anyone has a great solution to date.
>
>
> -----Original Message-----
> *From:* Stephen Wingfield [mailto:steve at bicom.us]
> *Sent:* Friday, July 28, 2006 4:14 PM
> *To:* Asterisk Users Mailing List - Non-Commercial Discussion
> *Subject:* Re: [asterisk-users] Looking for carrier grade redundant
> solution
>
> Chris,
>
> Heartbeat failover will usually be your best mixed approach.
>
> As always there is a cost benefit to be considered.
> Where the call absolutely has to stay up then Fault-Tolerant software and
> hardware is the only option that works with Asterisk to date.
>
> If however you wish to keep costs to a mimimum then possibly an onsite /
> hosted model where the back up is available remotely. This model depends on
> set up however.
>
> In all cases I would suggest you take a peruse of PBXware : http://www.bicomsystems.com/products/online_demo/
> which is our SMB Edition. We will next week launch our Call Center
> Edition that is packed with features and functions to assist the running of
> a dedicated to running a Call Center efficiently.
>
> Feel free to contact me offline steve {at] bicomsystems {dot] com and can
> make more precise suggestions according to requirement.
>
> Regards
> Steve
>
>
>
> ----- Original Message -----
> *From:* Christopher Aloi <chris.aloi at gmail.com>
> *To:* Asterisk Users Mailing List - Non-Commercial Discussion<asterisk-users at lists.digium.com>;
> asterisk-biz at lists.digium.com
> *Sent:* Friday, July 28, 2006 3:44 AM
> *Subject:* [asterisk-users] Looking for carrier grade redundant solution
>
> Hello List -
>
> We are looking add Asterisk to the core of our voice/data network. Our
> first application will provide a hosted call center application for a number
> of tenants (customers) who will have between 5-20 agents (seats) answering
> ingress calls. The calls will ingress and egress the Asterisk server SIP
> (all TDM is handled by Sonus switches).
>
> My goal is to design a redundant solution using a multiple Asterisk
> servers with an NFS mounted filesystem.
>
> I've done some reading regarding Asterisk redundany, and so far it seems
> the best approach is running redundant hardware (power supplies etc),
> matching servers (with a heart beat ping between them) and a NFS filer for
> storage (hot swapable) connected to each box via gigE.
>
> Am I on the right track? Any other suggestions or resources I might have
> missed regarding developing a redundant solution?
>
> Thanks for your time,
>
> _Chris_
>
> ------------------------------
>
> _______________________________________________
> --Bandwidth and Colocation provided by Easynews.com --
>
> asterisk-users mailing list
> To UNSUBSCRIBE or update options visit:
> http://lists.digium.com/mailman/listinfo/asterisk-users
>
>
> _______________________________________________
> --Bandwidth and Colocation provided by Easynews.com --
>
> asterisk-users mailing list
> To UNSUBSCRIBE or update options visit:
> http://lists.digium.com/mailman/listinfo/asterisk-users
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-biz/attachments/20060729/4668c550/attachment.htm
More information about the asterisk-biz
mailing list