Portland Perl Mongers: Relational DB vs Key-Value Store and Beyond
Most classes require prior sign up. Please check in at the front desk when you arrive to let them know you are here for the class. Bags must be checked at the front entrance.
Panel: Selena Deckelmann, Igal Koshevoy, Jeff Lavallee, David Wheeler
This will be a panel discussion about the ups, downs, ins, and outs of relational, row, key-value, and hierarchical data stores (simplistic buzzwordiness: SQL vs NOSQL aka ACID CRUD.)
The panel will discuss parallelism, scale, data integrity, normalization, business logic, ORMs, and performance. Some of the following might be addressed:
- why do you want a relational DB?
- why do you not want a relational DB?
- Membase, MongoDB, Redis, Cassandra
- Tokyo Tyrant, CouchDB, old school K/V (zodb, bdb)
- distinctions between "relational" and "row store"
- how filesystem settings affect the database
- how important is your data?
- common errors in SQL schemas or usage
- is count(*) supposed to be fast?
- efficiency vs speed vs parallel cleverness
- what is "scale" and do you need it?
- massively denormalized, or massively normalized?
- ORMs, materialized views, indexes, and the query planner
- typical performance with small/large, simple/complex data sets
As always, the meeting will be followed by social hour at the Lucky Lab.