I always thought it's the app itself calling mobileiris (negates
the need to maintain a proxy server), because for a proxy to cache
all service-stop pair results every few minutes is going to bog
down the cache server, esp if usage was never high.
But I agree that the steep rise in mobile internet &
smartphone prevalance has contributed to increased load on the iris
server that should plateau when all smartphone users have access to
it.
In other citiess, value-added services like this is seen as an
example of good customer service and going the extra mile. cost of
maintaining a server means nothing if it meant you can provide a
service better than what your competitors can offer. interestingly,
more and more city governments are taking this task upon themselves
as many of the networks are run by many small and medium bus
operators with no common interface. let's see if LTA would take a
step towards extendimg arrival timing boards to every bus stop,
because even a single-line scrolling LED is good enough for a bus
stop with about 6 services each displaying arrival time for
5secs.
Ah, not really. For sgbuses, the developer has his own server to
cache the iris queries. Sbst's app too uses a special web page to
get it's results from too instead of its own mobileiris page.
A powerful database server to generate the timings can easily
cost in the 5 digits region, and that is not a small sum since sbst
already provides a better service than the other company can... To
pay that amount to make it slightly better is not really
justifyable. Hope smrt will get theirs running, so sbst will make
theirs better or the other option is to give a good optmisation to
the system... Its not a small sum to setup the infrastructure for
such services, gps modules(which both already has), mobile data
modules(which both already has too), servers, software and monthly
data charges and power is just some of the things they have to
pay.
