Cloud and NoSQL have a Genetic Relationship
Attendees can expect to have their version control world changed!
At JAXconf in
San Jose next month Tim Berglund (August Technology Group) will
give two talks on NoSQL, as well as a Git tutorial with Matthew
McCullough. We spoke to Tim and asked him about the current state
of NoSQL, and what attendees can expect from his sessions.
JAXconf: In your opinion, how is the rise of
cloud computing impacting on the NoSQL movement?
Tim Berglund: There is definitely a genetic
relationship between the cloud and NoSQL, and there is ongoing
synergy besides. Two NoSQL products actually have their origins in
cloud plays. The company Basho actually started out with the
mission of creating Salesforce.com apps. They knew they were going
to need a data store they could count on to perform at scale, so
they implemented a Dynamo-based distributed hash table. Their SaaS
business didn’t really take off, but they found that the
marketplace was interested in their database, which became the
product now known as Riak. Likewise, after exiting Doubleclick,
Dwight Merriman and Eilliot Horiwitz had this brilliant idea to
build a platform-as-a-service (PaaS) that would provide
functionality similar to what we now see in the Google App Engine
or SpringSource’s CloudFoundry. (This was quite prescient in 2007!)
Their PaaS offering never really took shape, but they discovered
people loved the database they built for it, which is now known to
the world as MongoDB. Now, “cloud” can mean a lot of things, but as
we look at the evolving continuum of infrastructure-as-a-service
and PaaS offerings, we find that relational datastores aren’t
really ruling the roost. Amazon has SimpleDB, the Google App Engine
has the BigTable-based Datastore, CloudFoundry has MongoDB. Each of
these has existing or planned relational offerings as well, but
they are at best peers to the NoSQL options, not the dominant older
brothers they have been elsewhere. So cloud has given birth to some
NoSQL products, and it seems like cloud economics are nudging us
toward NoSQL in subtle and not-so-subtle ways. In addition, cloud’s
progressive technological sensibilities make NoSQL adoption more
natural, compared to the conditions that prevail in a conservative
enterprise data center.
JAXconf: In your ‘NoSQL Smackdown!’ session, you will look at
some of the new technologies users must introduce to address
today’s scaling challenges. What technologies do you advise NoSQL
users to currently look at?
Tim: You should look at as many as time allows.
This is the period of early adoption for NoSQL products, which
means technology innovators who want to realize competitive
advantage by moving first are going to take risks on new products.
This requires developers and architects to do a lot of learning
about a diverse portfolio of products, and about a bunch of
computer science we used to be able to ignore. It’s so hard to
narrow the list down. There are so many good products out there,
and any time I leave something out, I risk leaving out a successful
solution. Developers and architects should really learn about as
many different solutions as possible, but since time is limited, I
will narrow that down. I recommend you study Cassandra, MongoDB,
Neo4J, Redis, Riak, and CouchDB. This is a good cross-section of
the data models, scale paradigms, and licenses you’re likely to
find in the NoSQL world at present.
JAXconf: In ‘Radical NoSQL Scalability with
Cassandra’ you will deep dive into the Cassandra database. Who
would benefit from attending this session?
Tim: Anyone who’s interested in NoSQL beyond
the level of the survey should definitely come to this talk.
Cassandra is just one option in the increasingly crowded NoSQL
field, but it’s got a very compelling combination of license, data
model, scaling paradigm, and community. Cassandra really shines in
“web-scale” applications, but it’s very easy to deploy in smaller
systems as well, and its growth path is second to none. If you’re
interested in how NoSQL database work under the hood and how they
can achieve true horizontal scalability, you should come check it
JAXconf: You will run a ‘Git Going with
a Distributed Version Control System!’ tutorial alongside
Matthew McCullough. What can attendees expect from this
Tim: Attendees can expect to have their version
control world changed. There’s no longer any doubt about the
importance of Git or that it represents the future of version
control. The community is embracing it, and the kind of
capabilities it brings us relative to previous solutions are truly
compelling. The only problem is that it’s different enough from
previous solutions that it can be a little bit challenging for
developers to pick up without adequate preparation. This tutorial
is exactly what you need to get over that hump. Matthew’s Git
training materials are quite polished and very effective. I’ve
delivered them with him before, and we have a lot of fun helping
people get started with this technology. Attendees will learn the
basic mechanics of using Git as well as how to reason a little bit
about Git internals and repository history. You’ll go back to work
as the go-to person for this exciting piece of technology.