If NoSQL is your answer, you are probably asking the wrong question.

If NoSQL is your answer, you
are probably asking the wrong
question.

Hi, my name is
Lukas Kahwe
Smith

For the conference we're
looking for someone who can
be a part of the 'Databases -
short talks' presentations to
give a lightning talk about SQL.

Given that most relational
databases have an SQL
interface

Key-Value-Store + Complex Queries
= NoSQL ?

I have created a database API
for unstructured
hierarchical documents
on top of an RDBMS with
an SQL-like interface

And I wasn’t even the ﬁrst
person to do it ..

Can we agree to stop using
the term NoSQL ?

Using NewSQL is not any
better!

Instead talk about RDBMS,
Doc-Stores, Key-Value-
Stores, Graph-Databases ..

There is clearly a growing
number of people considering
alternatives to RDBMS

Actually Key-Value-Stores
have been popular for many
years already

So clearly the current trend is
not just about sharding
unstructured documents in
an elastically scaling
cluster

Its also not about getting rid of
SQL, after all the CAP theorem
doesn’t talk about what
query languages to use

We still love RDBMS for what their good for
Storing and retrieving
structured relational small to
large data sets with high
consistency and reliability

And yes we also still love SQL
Running ad-hoc search
queries and schema
updates even if we are not
rocket scientists

http://2012.nosql-matters.org/cgn/slides/#olaf_bachmann

And while JSON queries might
be easier parsed by computers
SQL is deﬁnitely easy to
parse for humans

So what do we really worry
about .. ?

ACID, Asynchronous, Atomic
Updates, BigData, Binaries, CAP
Theorem, Column Oriented,
Elastic Scaling, Eventual
Consistency, Failover, FullText
Search, Graphs, Low Latency, In
Memory, MapReduce, Master-
Master, Sharding, Smart Clients

ACID, Asynchronous, Atomic
Updates, BigData, Binaries,
CAP Theorem, Column
Oriented, Elastic Scaling, Eventual
Consistency, Failover, FullText
Search, Graphs, Low Latency, In
Memory, MapReduce, Master-
Master, Sharding, Smart Clients

RDBMS vendors got stuck
as a result of their own
success

“The two biggest issues that stand out to me are
that current leading relational databases never
solved scale out very well, and online
operations are too expensive. [..] The
innovation hasn't been in the language, but in the
design of database engines themselves. “
Brian Aker (Drizzle, previously MySQL)

http://blog.krow.net/2013/03/mysql-vs-nosql-vs-postgres-vs-sql.html

The good news is that
RDBMS vendors have been
unstuck

However its hard to move on
from a monolithic
architecture, Drizzle is a
radical attempt at trying it

In many ways its easier to
start from scratch

But lets look at some
examples of things that have
happened in the RDBMS
world

VoltDB, ScaleBase and NuoDB
promise elastic scaling on
top of an RDBMS with SQL
and ACID* compliance
http://voltdb.com/tao-volt/ﬁve-principles.php
http://www.scalebase.com/solution/
http://www.nuodb.com/explore/sql-cloud-database-how-it-works/

Both PostgreSQL and
MySQL have native support
for JSON (and XML)

PostgreSQL performs reads
on par with MongoDB,
especially for larger data sets

http://www.slideshare.net/stormdb_cloud_database/postgres-xc-askeyvaluestorevsmongodb
http://jathanism-event-notes.readthedocs.org/en/latest/scale11x/talks/
postgres_as_a_schemaless_db.html

MySQL Server and MySQL
Cluster allow by passing SQL
via the Memcache protocol

http://blog.ulf-wendel.de/downloads/nosql_in_mysql.pdf

MySQL and PostgreSQL
both have forks providing
Multi-Master replication

http://www.enterprisedb.com/products-services-training/products-overview/xdb-
replication-server-multi-master
http://www.codership.com/content/using-galera-cluster

OQGRAPH Engine to store
graphs in Maria DB

http://openquery.com/products/graph-engine

MySQLND plugins for
client side query caching,
transaction aware load
balancing
http://php.net/mysqlnd_qc
http://php.net/manual/en/book.mysqlnd-ms.php

SQL is not a scalability enemy

How big is your data ..
really?

Operational experience only
comes over time, having
something solid that just
works matters

There can be innovation in
the RDBMS world

So is using Doc-Stores, Key-
Value-Stores, Graph DBs

Just pick the right tool for
the job ..

If NoSQL is your answer, you are probably asking the wrong question.

Recommended

More Related Content

What's hot (20)

Similar to If NoSQL is your answer, you are probably asking the wrong question. (20)

Recently uploaded (20)

If NoSQL is your answer, you are probably asking the wrong question.