Architecture and Understanding of MCU

[nomansaeed]

Dear All,

I am very very greatfull if you can help to learn me as well as other who wants about the Architecture of MCU especially about its Capacity, Back Plane, per port/site support.

How can these calculation calculated ?

NOMAN

[Sean Lessman]

Dear All,

I am very very greatfull if you can help to learn me as well as other who wants about the Architecture of MCU especially about its Capacity, Back Plane, per port/site support.

How can these calculation calculated ?

NOMAN

Which MCU?

Sean

[nomansaeed]

Dear Sean,

I read very carefully your views in different forums specially at Wainhouse. Thanks for keeping all of us updated.

Sir, my last message is not specifically for some MCU, it was for a general architecture.

However i am keen to know the architecture/capacity calculation for MGC/RMX and Codian 4500/8000 series. Also is it true that currently there is no Blade available for HD endpoint support for 8000 codian.

thanks for posting even on holidays...

noman

[Sean Lessman]

However i am keen to know the architecture/capacity calculation for MGC/RMX and Codian 4500/8000 series.

All MCUs are different. For instance, the TANDBERG MPS provides all features on all ports, but will drop port count as the speeds go up -- the capacity is specified at 384kbps.

The TANDBERG Codian boxes provide all features on all ports up to the full bandwidth of each port (no dropping of capacity due to speeds). This approach means the port count is fixed regardless of resolution -- other competitors would argue this is not flexible and should allow for more ports if the resolution is only CIF. Who is right? The answer really depends on the customer and what level of hassle and complexity (having to calculate capacities on a daily basis vs. never) they are willing to put up with.

The MGC capacity is affected by features. Because it is a chassis base systems that has a number of different cards available depending on the features you want. Additionally, the capacities of those cards are typically calculated at 384k or 768k and will drop as you increase the speeds. Another example is the audio cards have twice the capacity if used for telephone calls than if they are use for video calls. Encryption halves the port count and adding features like Continuous Presence or ISDN requires additional cards that take slots up that could be used for 'port' cards so the max capacity drops due to those features. The MGC is pretty complex to configure and calculate ahead of time the capacity on a day to day basis.

The RMX provides features on all ports to a degree, but as the resolution increases the max port count drops. They will call it resources instead of ports but this is all just marketing. The reality is the RMX is capable of 80 ports (when transcoding, continous presence, or speedmatching) of resolutions up to CIF, anything higher drops the port count 75% down to 20 ports. If you only care about voice switched (no transcoding, continuous presence, or speed matching) you can get up to 80 ports all the way to 'HD' (although this is HD out and not in, so some would argue as I would that this bridge isn't really an HD bridge). Keep in mind that using voice switched mode requires every endpoint in the conference to have exactly the same properties (speed, video algorithm, audio algorithm) or you connect as an audio only call -- not very flexible.

Didnt include RADVision/Cisco but let me know if you want me to. They have a complex model too similar to the MGC with a bunch of considerations depending on speeds and resolution.

Also is it true that currently there is no Blade available for HD endpoint support for 8000 codian.

Its coming. Wll probably be available before the other guys upgrade their hardware to do HD in and out. More details contact your TANDBERG representative.

Hope this helps.

Sean