[hydra-dev] Proposed Supported Platforms

[Jared Smith]

On Tue, 2010-06-08 at 17:55 +0200, Olle E. Johansson wrote:
> So tell me again why this is not Asterisk 2.0?

Please don't consider me the "definitive guide" on the topic, but here
is my (albiet limited) understanding:

First of all, let's talk about what Asterisk is and isn't

Asterisk:
 o  is a somewhat mature product
 o  is very feature-rich
 o  is typically deployed as a PBX (but can do much more)
 o  is not perfectly suited to distributed/federated architectures and
large-scale systems.

If we were to build Asterisk 2.0, I think it would be very much focused
on being a better PBX.  It would probably involve continuing the work
that's been done over the past couple of years (new bridging API, new
dialing API, better infrastructure for managing objects, etc.)

In my view, Hydra is different from Asterisk (and an Asterisk 2.0) in
these regards:
 
Hydra:
 o  is a technology preview, and will be for the short term
 o  is designed from the beginning to work in a distributed/federated
manner, with multiple fragments (for lack of a better word) on multiple
machines working together to contribute resources to the collective.
 o  is designed to scale much better than Asterisk ever could
 o  is most likely overkill for a small PBX installation (and maybe even
medium-sized 
 o  is not initally focused on being feature-rich, and may depend on
Asterisk (and/or other pieces) to provide rich features, until such time
as said features are built natively into the Hydra architecture.

One analogy I've used in explaining my view of Hydra is to look at the
SIP protocol.  As you very well know (I learned it from you!), the SIP
protocol is really just a set of primitives (INVITE, REGISTER, etc.)
that allow endpoints to communicate and agree on establishing sessions
between endpoints.  I see Hydra as being somewhat similar -- it's a set
of primitives that communicate with each other and allow for dynamic
discovery of new resources, all in an effort to get the call from point
A to point Z, and using all the necessary resources (routing,
transcoding, bridging, logging, fault detection, etc.) along the way.
We're not just building a better PBX here, we're building an
*infrastructure* that scales way beyond what a PBX can do.

Let me sum up my views with one more (poor) analogy.  Let's say we're
all Star Trek fans, and Asterisk is the Federation.  If we were going to
do build Asterisk 2.0, we'd build more ships -- bigger ships -- faster
ships and hope to make new alliances with superior races in an effort to
make a stronger Federation.

Instead of doing that, Hydra represents the Borg.  The Borg might take
over a few Federation ships to do its bidding, but at the end of the day
the Borg way of doing things is fundamentally different than the
Federation way.  With the Borg, you don't just have individual ships...
you have an intelligent collective.

--
Jared Smith
Manager of Poor Analogies 
Digium, Inc.