Article first published as What a Flock of Seagulls and Next Gen Cloud Computing Have in Common: Interview with Barry Morris, CEO of NUODB on Technorati.
Technorati had the opportunity to speak with Barry Morris, CEO of NUODB, at this week’s All About the Cloud Conference, hosted by the Software and Information Industry Association (SIIA).
His company is focused on solving elastic scaling of OLTP systems on the cloud. Unlike most other databases, NuoDB was built for the cloud from the ground up on an Emergent Architecture – a shared nothing, asynchronous, peer-to-peer approach that is ideal for the cloud, yet it delivers the power, reliability and functionality of a traditional database.
NUODB studied the most successful distributed algorithms, systems, and techniques of real world systems that provide scalability of service without proportional costs. They’ve taken those concepts and recombined them into a newly innovated client/cloud system that supports all the features necessary for client/server RDBMS systems. Meaning, the company’s solution is specifically architected to scale effortlessly on the cloud, without compromising on any of the features or guarantees of relational databases.
One of the key themes from the cloud conference was a reminder to not lose sight of the business challenges you are trying to solve first, followed by the technology to do it. “It’s not technology, it’s business,” one panelist shared. Along those lines, Barry identifies five key elements that are necessary for a truly open, effective cloud computing experience.
1. Elastic Scalability
This is where the bird analogy comes in. In the NUODB database model, no one computer manages, or is “in charge” of the others. You can add a low-cost node to the system to increase performance, but at the same time, removing it, doesn’t collapse the system. It’s a completely peer-to-peer system. That type of flexibility allows businesses to scale out or in according to their needs, without massive capital outlays.
“The average utilization of a database computer is between 5-10%,” Barry cites. This means there’s a whole lot of overhead sitting idle. That’s a business problem, and one NUODB wants to fix. with the capability to process more data with fewer systems. Because the NUODB system allows different databases to overlap and exist on the same machine or even run across multiple machines, system utilization is optimized better.
3. 100% Uptime
Not 99.99%. Downtime is a business problem, and another one NUODB ends because its one database, running in multiple data centers at the same time. “If Godzilla rises up an eats one of your data centers, that’s okay, because the system keeps running in one of your other data centers,” Barry explains. There’s no fail-over, wait, or disaster recovery needed.
4. Geo Distribution
5. “No Knob” Administration
“The human cost of database administration is 50-60% of the total cost of a traditional solution,” Barry estimates. That’s money that could be invested in product development or other forward impacting business expenditures. The usability and efficiency of NUODB’s solution aims to change that.
A lot of what Barry claims can at first pass sound like science fiction, and admittedly, I’ve barely scraped the topic. Fortunately, NUODB has a terrific webinar on its product that will take you data-hungry readers even deeper into the solution.