Wednesday, August 19, 2015
03:00 PM - 03:45 PM
|Level: ||Technical - Introductory|
Our development team set out to build an advanced Computer Vision platform and discovered we had several data problems to solve and architect around. While it would have been nice to find that ONE magical tool that solved all of our problems its hardly the case that teams can have that happen.
We ended up settling on an platform design that leveraged the "best tool for the job" from the open source world including:
Design considerations included:
- network latency in a public cloud hosting environment
- as close to zero downtime during maintenance/scaling windows
- picking the right tool(s) between storage/search of data
In this session we will talk about how we investigated various design options, including some pitfalls along the way and making changes to evolve our platform to allow for a greater level of flexibility moving forward. We will also cover operational considerations and tooling to allow our team of developers to manage dev/test/production environments will limited IT Ops support.
Nate has been working in the Enterprise and Mobile Software industries for the past 15 years in various capacities, from Developer, to Architect, to Product Manager. He is an avid user and advocate for Open Source technologies and is a PMC member of the Apache Software Foundation Bigtop Project that aims to set standards for packaging, testing, deployment of various Big data and NoSQL technologies in the Hadoop ecosystem.