MongoDB World Interview With an IoT Architect
A talk with Nick Brown, Principal Software Architect of Vivint of architecting a NoSQL solution for massive IoT with a focus on smart homes.
Join the DZone community and get the full member experience.Join For Free
Vivint is overhauling the smart home industry with its work on the connected home. In addition to working on a connected home (connecting all devices within), they also have their own offerings.
At MongoDB World, they talked through the challenges and triumphs of a high-risk, ultra-urgent, ground-breaking Internet of Things implementation. They quickly became device independent and managed to operate on a very fast level.
They have a mature MongoDB-based infrastructure that is two-and-a-half years in production with 600,000 systems installed and 500 million messages a day. That's 50,000 per second utilizing a MongoDB cluster of 50 nodes (30 nodes in one datacenter and 20 in another) and 10 shards. They also run machine learning on top of MongoDB and dump cold data to HDFS.
Their architecture is real-time and multi-tenant architecture in MongoDB, which manages stateful and historical data at scale.
In the smart home industry, there's a combination of hardware, software, embedded software, and platforms; combined, they require low latency.
They like MongoDB for its ease of implementation, flexible schema, and ability to handle new devices all the time. Working in the smart home industry, those sorts of capabilities are essential to their business. Not only can they add hardware and software as they change, but they can keep those changes as fields come and go.
Opinions expressed by DZone contributors are their own.