Re: [Legba2013] PapaLegba 2013 review.

[tim p. <th...@we...>]

You are right - I meant that we had too few towers for what we were trying to do.
If we want to cover the playa with 2 towers, we need more channels per tower, or
more towers. 
Alternatively we'd need to de-scope and only a smaller area.

T.

On 10 Sep 2013, at 15:49, Johnny Diggz <jhi...@tr...> wrote:

> It is significant to note, we had significantly fewer technical hands than we did last year.   Our fewer towers was not lack of effort, rather time.
> 
> 
> On Tue, Sep 10, 2013 at 6:52 AM, tim panton <th...@we...> wrote:
> 
> I've just spent the day working on the stats and I thought I'd write a note on
> my thoughts for the PapaLegba crew and my Tropo sponsors.
> 
> http://gont.westhawk.co.uk/bm2013/stats/report.htm
> 
> We did significantly worse than last year. Fewer calls, fewer minutes, fewer hours up.
> 
> We did provide a useful service - e.g. 80 5 min long calls that wouldn't have been made any other way.
> 
> We also managed to gather some raw data to analyse - which is helpful - thanks Mike!
> 
> What went wrong?
> 
> 1) We set up late. by the time we went live there were already lots of handsets to register.
> 2) The SMS delivery mechanism lacked the throughput to deliver the registration load, leaving no capacity for normal traffic. Even when it had been re-written (twice) on the Playa it barely kept up. When I added 2000 sms's notifying of rain on Monday we managed to only deliver 61 of them by the time we tore down.
> 3) We lost power _every_ night. The meant that any phone still powered up had to re-register every morning - even if it hadn't moved. This consumed much of our radio capacity.
> 4) We had too few towers.
> 5) The design kept changing - despite the promise that we'd test everything at RELIEF, use a VPN and use IMSIs as caller IDs - we did none of the above. 
> We rocked up late, with an entirely untested core network which was still being re-written on thursday. It is frankly amazing that anything worked at all.
> 6) We broke our promise to Ice to deliver a gate SMS application through to monday evening.
> 7) we didn't communicate our purpose very well to other burners.
> 
> What worked?
> 1) It looks like the new L3 held up ok, although I have no stats on that.
> 2) The ability to script the core network meant that we could re-write the SMS delivery (at least to some extent) so that it prioritized recent SMSs thus keeping a small number of people happy at least.
> 3) the interface with tropo worked - the ability to access a database from tropo was very handy for sms groups etc.
> 4) outbound voice and SMS to the PSTN
> 5) The inter-base station 3.5GHz network
> 6) PlayaNet as backhaul worked ok for us this year - perhaps becuase we didn't require anything special of it.
> 7) The provision of wifi and a hardphone was popular again.
> 8) group sms
> 
> 
> Overall, I'm doubtful I'll do this again, it will be difficult to justify the time spent for the results unless we make significant changes in our approach.
> 
> I'll write a separate note on the viability of scripting a core network in javascript.
> 
> Tim.
> 
> ------------------------------------------------------------------------------
> How ServiceNow helps IT people transform IT departments:
> 1. Consolidate legacy IT systems to a single system of record for IT
> 2. Standardize and globalize service processes across IT
> 3. Implement zero-touch automation to replace manual, redundant tasks
> http://pubads.g.doubleclick.net/gampad/clk?id=51271111&iu=/4140/ostg.clktrk
> _______________________________________________
> Openbts-bm2009 mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/openbts-bm2009
> 
> 
> 
> 
> -- 
> Chief Evangelist, Tropo
> Tropo.com || Geeks Without Bounds || Diggz.org
> @johnnydiggz || di...@tr...