[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [school-discuss] FOSS Professional development courses for teachers



--- JTD <johndenny@xxxxxxxxxxxxxx> wrote:
> Funk v1 is very cool too...   I dare ask what is the feature set?   :)

Dare I answer? I think I might get kicked and banned.
 
> Do you have any way we can log in and check out Groove...  as far as I can
> see by its feature set it is nothing special...  just a cool name...  but
> you never know till you jump in and give it a try...   net connections are
> too slow here in Cambodia for me to explore too much...  wish someone would
> do a comparison chart of various school collaboration tools available.

To me Groove is just vapourware[1]. I've never seen it operate. My
understanding of its operations comes from:
 (a) Marketing material.
 (b) Presentations made by the principal driving the project (the school only
has 2 or 3 production sample devices).
 (c) First-hand accounts related to me by other LUG members who *have* used
Groove in real-world work.

Items (a) and (b) are pretty much in lock-step so I think Mr. Principal is
quoting the serving suggestion on the back of the box.

My appologies, this is about to get pretty technical. You might want to skip
this bit.

Combining (c) and [2] (below) and extrapolating based on (a) and (b) leads me
to believe (and this is *entirely* opinion) that the following is likely to
occur:

 * Initial delivery of no more than 30 devices will occur (we'll get to this
number below).

 * Initial trials will show promising results, with some teething problems,
lost data, license issues, insufficient memory and slow response. Not enough to
cancel the project.

 * Based on initial roll-out the school will commit to purchase 300 more.

 * As the rest start to be rolled onto the network, the cracks in the UDP based
synchronisation protocol show up and the N^2 nature of the ad-hoc networking
will start to take its toll on performance. Layer 2 collisions will start to
outweigh useful traffic (classic CSMA-CD utilisation curve).

 * After about 120 devices are added to the network, the IT guys will be
pulling their hair out and eventually dividing the system into smaller and
smaller separate networks to try and work around the inherent bottlenecks in
the network topology. But this will defeat the purpose of the ad-hoc networking
and the peer-to-peer apps that were bought. It will also mean that 3 maybe 4
times as many WLAN access points will be needed to overlay the multiple layer 2
to layer 3 network edges onto the school campus area.

Based on what I've heard about using Groove, it works ok in small office
scenarios below about 10 machines, but as the number of devices grows the
synchronisation between devices becomes slower and slower.

So (aherrrm) in the best interest of the school, the supplier will graciously
agree to a small trial of say (how many units before the scalability issues
kick in?) 30 units before requiring commitment in full for the remainder.

I'm such a cynic!


--
  Rich

[1] Queen's English here.
[2] My day job is as System Architect and Principle Engineer for a wireless
communications R&D company supplying equipment and networking expertise to
financial institutions across Australasia and SE Asia. The ATM network in
Australia for which I designed and wrote the majority of the server component
is now at uptime 528 days - it's had zero downtime since commissioning (thanks
RedHat). Yeah, sorry, that was narcissistic; but I'm proud of that system.



      Get the name you always wanted with the new y7mail email address.
www.yahoo7.com.au/y7mail