[Ongoing series of blog posts to inform
potential developers, users and (hopefully investors) about this new app
ecosystem I am architecting, designing, developing and deploying. More details
at this page]
A lifetime of working with computers has taught me a couple
of things, and one of the thing I learnt is the amount of issues that are
raised because of poor supply in, well, in the cities we live in. For reasons
that are beyond the scope of this document, it is essential to focus a lot on
backup systems, and power management is the first big topic on backups.
One must assume from the word go that we cannot always rely
on continuous power supply, where we live. Power cuts can go from 10 minutes to
an hour to 8 hours to a day. It is simply impossible to assume that something
will be available without interruptions. That means, if an essential project is
at hand, then, it is best to prepare for every eventuality.
That is why having multiple computing devices (as explained
in the hardware section) with at least some of them powered by portable battery
becomes essential to keep work proceeding without any issues. That is why, I
recommend at least two battery powered computing devices, and not to forget
general backup supply to the main workstation PC.
Further, it would also be a good idea to invest in a home
power backup system, although, I must admit that those things cost a bomb. If
the task at hand is critical enough, investing in a diesel power generator is
also a good option, and if the supply of money is good, even essential.
Computing devices can break down for a number of reasons.
They could get stolen. Parts could simply stop working, and input devices in
particular can die without any warning. All in all, redundancy takes on a brand
new meaning because of this.
That is why, multiple computing devices must be hand. Also,
don’t skimp out on the quality. If necessary, pay more for less (buy lower
spec’d machine but go for the expensive warranty) so that what you have will
work no matter what.
Developed code must be backed up in multiple online
locations, in servers located all around the world. Offline backup must be made
multiple times, and must be stored in multiple locations. Obviously, backup
restore must be tested, and disaster plans must be exercised and recovery plans
tested and verified when the worst happens.
A recovery plan (tied to a backup plan) that does not work
means, you are effectively walking around without a helmet.
Redundancy also means that all the work (at least the one’s
that are currently being worked on) are cloud enabled. That way when disaster
strikes, you should be able to setup a completely new work machine into your
own, assuming superfast internet access. This is perhaps, the most essential
redundancy plan there should be.
When it comes to backup and redundancy, a certain amount of
patience and planning becomes essential. The idea behind having a backup and
redundancy plan is not because you expect something to go wrong. In fact, many
projects go just fine without any issues. That makes the entire activity
related to backup and redundancy an exercise of wastage. This is similar to how
some people don’t want to pay for warranty extension because, well, most of the
time, the hardware simply works. However, errors are part of any system and
such a system manufactured the hardware. So, yes, having warranty makes sense.
Remember this. You might as well have something and not use
it, than need it, and not have it.
[Last Updated February 10th 2017]
Follow me on twitter,
for more updates. Thanks!