I'd heard previously about the 7 +- 2
rule. I heard that it meant someone could only do around 7 things at
once – though I wasn't told this was between 2-3 bits of
information. It certainly seems interesting that short term memory
isn't based on bits but by chunks – certainly encouraging people to
specialize instead of being a jack of all trades.
In regards to failures in cloud
computing, I suppose I previously always defined “uptime” as time
that the service is available and working correctly – though I
suppose this time frame may differ depending on the software. I think
the distinction between accessibility and availability is an
important one. Having a feature that presents a users with a
descriptive screen of what's going on when the service is not
available can be critical to keeping users content, happy, and
informed.
Degradation instead of failure is a
wonderful idea. Nothing is perfect and everything will fail
eventually – it's better to plan for these events. This also helps
encourage a healthier state of mind and development practices – to
not assume you're design is perfect. I've always felt that redundancy
of middle-tier equipment has had an edge over singular high-tier
equipment – it's less presumptuous and gives multiple failure
points. The points on being able to detect failures was basic but
important.
The point on making sure a “re-try”
is included in exception handling seems obvious but probably often
forgotten and critical. I hadn't previously thought of all the uses
of request-buffering though – I'll certainly try to remember it for
later projects.
Capacity buffering seems like a basic
good design, and dynamic resource accessing seems like a good idea if
even for security reasons. Automating responses to failures as much
as possible seems like a normal step as well – as does seeking out
problems and solutions which may pertain to your system.
After reading the title of the paper
on tire's and mandatory wireless service, I'm immediately concerned.
I don't particularly desire for my car to be communicating with
anything else other than me physically – that way I'm more secure
from being remotely screwed with. It seems rather obvious that a
particular network can be tracked with some work – while not a
particular concern for me at the moment – could easily be for
others. Another rather disgusting feature is the mandatory
requirement for it as opposed to letting it be voluntary.
I'm not sure if I should be proud or
concerned that some of the experiments done concerning privacy were
done somewhere I drive regularly. While I wasn't shocked to learned
that the range of picking up packets was large, and that it was not
difficult to force trigger and capture it – I was concerned. There
seems to be very little to no security on the system as well –
which is even more concerning. The later experiment that forced
warning lights to go off brought everything into perspective – you
could force someone to stop in order to render them more vulnerable.
No comments:
Post a Comment