You are viewing an old version of this page. View the current version.
Compare with Current
View Page History
Version 1
Current »
Overview
The performance of the SBC VM can be optimized if the call mix is provided while configuring the SBC SWe instance. A set of predefined call mixes , these are called standard traffic profiles. Additionally, can create custom traffic profiles, with a call mix that is SBC SWe.
The supported standard profiles are:
Supported Standard Profiles
Profile Name | Description | Minimum vCPU supported | Maximum vCPU supported |
---|
default | By default, non-cloud instances comes up with default profile. This profile retains the SBC SWe behavior of earlier releases (5.0.x and 5.1.x), where signaling and media vCPUs overlapped. | 4 | 24 |
standard_signaling_profile | This profile utilizes the vCPUs for signaling (direct media) calls and it is auto-configured on the cloud-based S-SBC instance. | 4 | 40 |
standard_passthrough_profile | This profile utilizes the vCPUs for passthrough calls. | 4 | 40 |
standard_transcoding_profile | This profile utilizes the vCPUs for transcoding calls. | 4 | 40 |
standard_callmix_profile | This profile supports callmix which contain a combination of 90% pass-through and 10% transcoding. | 4 | 40 |
standard_msbc_profile | The purpose of this profile is to utilize the vCPUs primarily for passthrough calls. This profile gets auto-configured on cloud-based M-SBC instance. | 4 | 40 |
standard_highcps_callmix_profile | This profile is similar to standard passthrough profile with higher CPS and the corresponding CHT is 25. | 4 | 40 |
standard_gpu_isbc_profile | CPU layout for this standard profile is pre-defined and optimized for 100% transcoding on an Integrated SBC with an assumed call hold time of 133 seconds. | 6 | 32 |
standard_gpu_tsbc_profile | CPU layout for this standard profile is pre-defined and optimized for 100% transcoding on a T-SBC in a D-SBC deployment with an assumed call hold time of 100 seconds. | 6 | 32 |
Once the distribution of vCPUs is set, the SBC SWe overload behavior functions the same way as the hardware SBC (5000/7000 series).
Caution
This feature is applicable only when the vcpu is more than or equal to four.
Profiles
Standard Profiles
When any standard profile is activated, the application automatically optimizes the vCPU usage of the SBC SWe VM for the activated profile.
Caution
- SBC SWe instances reboots automatically to activate the profile. All active calls are lost during this process, it is recommended to activate the profile during maintenance.
- A trap is generated when a new traffic profile is activated.
Appropriate profiles are auto-configured on the D-SBC based on the personality as follows:
Personality Types per Profile Name
Personality Type | Profile Name |
---|
I-SBC | default |
S-SBC | standard_signaling_profile |
M-SBC | standard_msbc_profile |
MSBC (GPU Transcoding mode) | standard_gpu_tsbc_profile |
MSBC (CPU Transcoding mode) | standard_transcoding_profile |
Table Name | Description |
---|
sweTrafficProfiles | Contains all the standard profiles and their fixed call-mix. |
sweActiveProfile | Contains the name of active profile and time-stamp of activation. |
sweCodecMixProfile | Contains the standard codec_mix profiles. - G711_20ms with 100% G711 (20 ms) calls.
- G711_G729_20ms which is initialized with 50% of G729AB (20 ms) and 50% G711 (20ms).
|
Table Name | Description |
---|
sweProcessorCapacity | This table contains the real processor capacity indices relative to a 2690v2 processor with the recommended BIOS/ESXi setting. |
sweCapacityEstimate | This table contains the capacity estimations for all the standard profiles. |
Custom Profiles
If the call mix is different from one of those corresponding to the standard profiles, a custom profile can be created. You can create custom profiles based on the call mix requirements. While creating a custom profile, you must provide the call mix through CLI.
Caution
The estimated calls per second (cps) numbers are calculated based on certain conditions. The actual traffic may have a different number of signaling messages per call, message size, or call features for which the estimate could be inaccurate.