On 2010-02-08 18:57, Brandon M. Lapointe wrote:
> I've been a "Cisco guy" for many years now, but (as I'm sure many of you
> also did) I started many moons ago as a "Microsoft guy". Microsoft did
> have various solutions for the lab/education/skills development arena
> such as the (already mentioned) MSDN, MAPS (action pack), and various
> Eval/Beta licenses. Now many of these models would not be applicable
> to Cisco's situation keeping in mind some of the legitimate business
> needs (such as generating revenue) that Cisco faces, but I think a
> Cisco licensing model could actually open the door for the kinds of
> solutions this community is looking for.
Sure. But let's face it - some part of software Cisco develops
(I don't know exactly, but I'm propably not in position to discuss it
with the community without NDA) is dependent on the hardware. Microsoft
(and some other companies, let's don't stick to this one company) is
doing software that runs as a whole on a readily available PC.
The problem with moving IOS (or any other hardware-specific OS) to
"masses" is always additional cost that is required to "build"
specific version of the code to use as a working PC piece. Dynamips
as wonderful as it is, doesn't support a lot of features IOS executing
underneath can - it breaks in tight time-coupled functions like
BFD for example, not to mention totally out-of-sync RTC which
agains hits in various places around the system.
However - the cost is already paid, as I said in previous mail -
Cisco is using some tools internally that are oficially builded
in a specific manner to behave correctly on PCs - both Intels x86
and Suns SPARCs. If you did hear IOU, you propably did hear IOL.
> For instance, if hardware had some kind of strong licensing mechanism,
> wouldn't Cisco be able to dramatically lower the price for some of the
> hardware since it would be useless without the license to "unlock"
> various feature sets that are then highly price dependent?
Nice thought, I haven't thought about that in this way, but on the
other hand - while most of the newer gear already has hardware
counterfeit and copyright protection elements that work with and
without connection to the software running on it, please remember that
the whole unit (meaning - board, various pieces, processors, ASICs, etc)
is usually quite complicated and not so cheap - no matter what you
think about Cisco margins.
So while in time the price ratio between hardware and software
will be more to the software part, the hardware itself is
becoming more and more specialized and thus - not so cheap.
There are some modules for ISRs for example that can be "emulated"
(read - cheated to think they run on a real module), because
the module itself in reality is "just a PC" but in a different
hardware form and with some Cisco-specific motherboard connection
and signalization. The design was copied (I won't mention names
but You can easily figure them out), but what is worse - it
was also copied to build counterfeit copies of the module.
So additional counterfeit protection was built in to the
solution and it will be hard for the software running on the
next gen modules to be 'hacked' the same way as before.
This is just one of the 'problems' vendor is facing when tries
to be as open as it can without loosing competitive advantage
over others.
> would open the door to a "training" license that restricted say
> number of interfaces and/or PPS throughput, which would accommodate the
> educational segment while rendering it all but useless to a
> production environment
When we speak only about control-plane that's propably doable,
and in specific - for IOS it was already done. If You think about
OSes Cisco uses in general - it's harder to judge. IOS-XE or NX-OS
runs on a Linux kernel, so in theory it's pretty easy to "port" it
to run under normal OS or within VMWare. IOS-XR on the other hand
runs on QNX and the stuff is so complicated it requires a lot of
hardware ($$$, trust me) to "emulate", so it's unrealistic to build
a whole network simulating a couple of IOS-XR-enabled nodes. At
least for now of course.
-- "Everything will be okay in the end. | #ukasz Bromirski If it's not okay, it's not the end. | http://lukasz.bromirski.net Blogs and organic groups at http://www.ccie.netReceived on Mon Feb 08 2010 - 19:51:57 ART
This archive was generated by hypermail 2.2.0 : Mon Mar 01 2010 - 06:28:35 ART