RE: [Embedlets-dev] Re: [Arch] ANT

[Christopher S. <cs...@oo...>]

>
> Topic tags:[Arch][Wiring][Doc][HW][PM]
> _______________________________________________
>

Ted asks:

>For example, if a change needs to be made
> to the code
> in the muivum devices, does someone walk around to 80 devices and manually
> upgrade each one?  Even if these devices were accessible through FTP, does
> someone manually FTP the code changes into each device?  What if all the
> devices need upgrading or if 3 more injection molding stations
> are added to the
> plant?   With Ant, one just types 'ant [params]' in the right
> context and any
> piece or group of pieces in the embedded part of this
> application's build and
> deploy process can be executed.

In the case of statically configured Embedlets or an Embedlet bug
fix/additions this would require the code base to be updated via whatever
means available. In the Ajile based systems it appears that a Jag physical
connection needs to be made, unless they come up with a flash bootstrap over
TCP/IP. Other systems may rely on ROM substitution. I do not know what
muivum provides but I would be (pleasantly) surprised if they have
remote-able flash update. I do not think that Ant would be of assistance in
overcoming these barriers.

Dynamically configured Embedlets would have the ability to modify settings,
rewire connections and even instantiate new Embedlets from the available
pool. Ant and other automation tools would be of great assistance in this
case.

This highlights another advantage of dynamic configuration in that once the
code base has been QA'ed and stabilized it would not need to be changed and
all devices of a similar function would share the same code, greatly
increasing reliability and reducing maintenance.