Plugging the traffic drain

DataStax Enterprise 4.0 goes GA with shiny new in-memory database

Lucy Carey

Commercial Apache Cassandra is here with souped up search capabilities, and a new solution to help with pesky attention deficit online customers.

, the company feeding Apache Cassandra to the
enterprise masses (400 customers to be precise, 25% of those
companies), announced the GA release of DataStax Enterprise
4.0 today – complete with a new in-memory database option. We
caught up with Product Team member Robin
Schumacher to find out what the implications of
this addition are for fans of super-scalable Cassandra.

In addition to this ‘headline’ feature, DataStax
Enterprise 4.0 also offers souped up search functionality, an
updated version of the OpsCenter visual monitoring tool, and
Cassandra 2.0

The most obvious advantage of inserting an in-memory
database to your NoSQL system is speed, speed, speed. We’re all of
us becoming increasingly impatient with slow applications, and
Schumacher reckons that the majority of people right now expect
everything they’re doing on a webpage to finish in two seconds or
less. It’s this attention-deficit mentality that’s driving the rise
of in-memory databases.

UK retailers lost 8.5 billion Euros because of
 slow websites in the last year

If pages fail to load within this micro-timeframe,
potential visitors will simply head to a different site. Citing a
study which revealed  that UK retailers lost 8.5 billion Euros
because of  slow websites in the last year, Schumacher notes
that, clearly, businesses want to try everything in their power to
stop this revenue drain. Thus, “in-memory becomes one of the tools
at their disposal to help with this, and that’s what we’ve now
added to our platform.”

He observes that what DataStax customers can now
do is to “really dial that performance, and target their data
wherever it’s needed in terms of application

Ultimately,  the goal for DataStax with this
release is to “take all of the goodness that people enjoy with
Cassandra today – the flexible data model, the multi-data centre
support, the linear scale out capabilities – and bring all of that
to an in-memory database.”

It’s not just in-memory computing that’s seeing a
spike in interest. Across the board, Schumacher is seeing customers
becoming more educated and aware of the NoSQL market. By offering
an enterprise class NoSQL platform that provides users with “the
same type of maturity and niceties that they’re used to from the
relational database world, coupled with the power and flexibility
of Apache Cassandra NoSQL”, DataStax is making strong inroads into
the burgeoning user base.

NoSQL is coming of age…and developers are changing
their approach to database planning

The company doubled in size last year, and plans are afoot
to do this again in 2014. NoSQL is coming of age, and along the
way, developers are changing their approach to database planning.
For Schumacher, customer feedback in the past was “almost a broken
record.” Again and again, people developing heavy duty online
applications  were  “trying  to shove the square peg
into the round hole by starting with a relational database. They’d
start with Oracle, MySQL, and all that, then they’d start to look
for a NoSQL solution.”

Now, however, Schumacher is seeing  customers who
are beginning with NoSQL.  He notices that, “They already know
when their app isn’t a fit for a relational database. It’s not a
why should I anymore – it’s how do I? How do I make this switch?
How do I make this happen? That’s the change I’m seeing. “

People are using NoSQL to make money, as opposed to
counting it

Schumacher started off as relational database admin,
and has noticed that  relational databases are very good at
what he calls “counting money.” So, “accounting applications,
things like that – that’s what they’re really suited for. In our
new world of online applications, that’s where NoSQL is coming into
play. People are using NoSQL to make money, as opposed to counting

A key  example of this is DataStax customer Netflix,
who run 95% of their data runs on software from the company.
 Schumacher can’t release exact statistics, but he can tell me
that they are the number one data generator on the internet right
now. Running completely in the cloud, across multiple availability
zones – all designed to ensure that no Breaking Bad marathon is
troubled with lags and downtime – Netflix is, to Schumacher, “ a
really good example of today’s modern online application that
benefits from NoSQL.”

DataStax remains staunchly open-source, and continues
to contribute over 80% of the enhancements and the code that goes
into Cassandra. Although they offer a separate community for users,
alongside the core, growing Cassandra community, there’s a good
deal of convergence between the two, with many devs criss-crossing
back and forth. As the major driver of the software, DataStax make
a point of extending things like free open source drivers and free
online training for the open source community at large.

Netflix is a really good example of today’s modern
online application that benefits from NoSQL

Although there are other NoSQL databases with
in-memory computing, what Schumacher believes sets DataStax apart
is the sheer scalability their software offers. As an example, with
one of their in-memory toting competitors, the largest cluster size
you can have is only 30 terabytes. In contrast, DataStax customer
Ebay have one Cassandra table that’s forty terabytes inside: Just
one database object, versus an entire database cluster.  

He proudly states: “It’s that type of scale that we
offer, and now we can solve those very fast in-memory use cases
like these other in-memory NoSQL databases do.“

comments powered by Disqus