Startseite › Forums › HAast (High Availability for Asterisk) › General › Backup PBX in cloud/AWS
-
AuthorPosts
-
We have an onsite PBX and we are considering creating a backup PBX on Amazon Web Services (AWS) or another cloud provider (e.g. Azure). Is HAAst compatible with AWS or Azure?
As well, will any underlying network protocols work over a long distance WAN (we’re not sure where AWS/Azure will actually host our PBX)?
Yes. No problem!
Many call centers are keeping their primary PBX’s on site (for performance and other reasons), and using the cloud for their backup PBX’s. HAAst will run on Amazon Web Services (AWS) + Amazon Elastic Cloud 2 (EC2), Microsoft Azure, and more, to create your primary PBX, backup PBX, or both. You can even place the HAAst operational database in Amazon Relational Data Services (RDS) or equivalent with both peers writing performance and operating data to the shared database. (The HAAst web GUI allows you to view combined and individual PBX reports in many areas). We do not recommend placing Asterisk configuration databases in RDS or equivalent – but this should not be a problem as they are fairly static and small.
Unlike other open source and closed source commercial HA products, HAAst does not use any LAN based protocols (NFS/DRBD/hearbeat/etc). HAAst is designed around only WAN based protocols, primarily it’s own PeerLink protocol. HAAst can accommodate large swings in network latency as would be seen in links between in-house servers and the cloud. HAAst (Commercial Unlimited edition) even learns these variations and adapts to the changes, so you never have a false positive fail-over of the cluster. This approach allows customers to even place their PBX instances in different Amazon Recovery Zone’s, and in different Regions.
With AWS, or any cloud provider, you will want to use their static network hardware option; Amazon calls it Elastic Network Interfaces (ENI). As of August 2016 Azure (ARM and ASM) support static network hardware (MAC address).
Since the above post was written 1 year ago, the market trend has reversed and we are seeing more companies move their PBX’s off the cloud and back on-premise. Reasons vary but we have frequently heard that audio problems (due to latency) are forcing the return of PBX to a hardware platform under full control of the operator.
On the high-end, cloud providers like AWS (even on their XL platform) are encountering latency problems at approximately 250 simultaneous calls. This number is not consistent but as of October 2017 we have heard from numerous large scale providers that AWS is not consistently reliable for large deployments. (Although we have a couple of HAAst users with well in excess of that number of calls on a single AWS instance). There appear to be variations by host/data center/region affecting performance.
Telium continues to support any combination of on-premise, cloud, and hybrid. Telium would be pleased to assist with the design of any telephony environment, building on the experience of a wide range of customers.
-
AuthorPosts
- You must be logged in to reply to this topic.