Nitai,
Many thanks for the great insight!
In response to your 4 items:
1. *CODE*: Very good insight and agreed. I do have a New Relic account
- Check!
2. *DATABASE*: I will check your links in a few... One thing that I've
seen consistently across many blogs through some minimal research in db
comparisons is that Mongo is easier to setup and get going on and even
though it can scale, Cassandra is easier to configure for scaling and setup
for scaling (for a cluster) by default (or something similar to that.) The
sharding stuff for Mongo seems to, again from the minimal research I've
done, even on the Mongo website, seems to be very "warning-like". They say
if you you don't plan from the start with how you are going to shard the
db, it can become a pain later on. Does anyone know of any very good
reference tutorial on properly preplanning/preparing for sharding with db
structure examples for Mongo specifically?
I also understand that several of the major high traffic sites use
multiple databases as each db has its own benefit in performance, ease of
use/management, scalability, ACID transactions, etc. My DB need for now is
for the user/categories portion of the apps. So far, you've given me an
extra star for Mongo ;)
3. *SERVER*: It will all be EC2-based probably using Beanstalk since we
don't support GAE anymore (because of GAE inconsistency issues). Been
hearing a LOT about Nginx performance all over the web. May definitely
consider this - otherwise apache.
4. *DNS*: Already have DnsMadeEasy - Check! They are very good by the
way - I've used them for years and they've recently done a major overhaul
with their UI and other features. I know that AWS has their Route53 or
something and one of the projects I was working on seemed to have made it
easier to work with a load balancing setup with their elastic IP. I still
like the decoupling from AWS with DnsMadeEasy though.
Nitai, could you elaborate a little bit more on "* You
can even achieve DNS load balancing and diverse network connections with
DNS.* " The part I'm talking about is the *diverse network connections*.
Do you mean for failover purposes or scalability. If for scalability,
please elaborate.
Your response has been very helpful. I may consult with you in the near
future as well if your prices are not anywhere close to what I "Think" a
guy like Alan's could possibly be ;) His prices "might" be more affordable
once I reach phase 9 or 10 of the project. I wonder what the heck his
consulting prices could be like????
Anyhow, now's the time I really regret not going to the open cf summit!
Didn't have the time to travel at that time... Would've been nice to hear
Alan speak on high performance OpenBD!
>
--
online documentation: http://openbd.org/manual/
google+ hints/tips: https://plus.google.com/115990347459711259462
http://groups.google.com/group/openbd?hl=en
