ADT Pulse Smart Home Security Scalable IoT Architecture Design and Operations Support

ADT Inc. provides residential, small and large business electronic security, fire protection and related alarm monitoring services throughout the United States. ADT Pulse is its high-end home security and home automation service, used by millions of customers, combining alarm and security service with home automation and live stream video.

The software system underneath ADT Pulse was built and operationally supported by iControl Networks, now part of Alarm.com. iControl engineered the application to store data in the Oracle RDBMS — configurable as multi-node RAC, optionally on Oracle Exadata engineered systems, and mirrored to DR failover sites in different geographic regions. iControl also developed a version that runs natively in the Cloud and can use DataStax Enterprise Cassandra as a massively scalable, cost-effective store for the IoT data the application captures.

As adoption of the home security product grew and total data sizes with it, RhinoSource advised iControl's engineering and IT support teams on database scaling, performance tuning and coding best practices — work that helped them win top cable and home security companies as customers.

To carry that growth, RhinoSource designed an Oracle Partitioning architecture for the application's database schema, deployed on fault-tolerant Oracle RDBMS clusters supporting millions of US subscribers continuously. We wrote the automation that runs daily to maintain those large partitions and drop expired IoT data without disrupting an always-on application, and revamped iControl's schema install/upgrade utility so new releases could be applied with the system online — a hard requirement for meeting strict customer SLAs.

Planning further ahead, for web-scale data sizes, RhinoSource educated the iControl team on NoSQL architectures and designed a high-performance Apache Kafka and DataStax Cassandra IoT architecture capable of capturing and storing hundreds of terabytes of IoT sensor data from millions of homes, cost-effectively. We designed the Cassandra schema and worked with iControl's engineers to port the application onto Cassandra for storage and Kafka for streaming, using simultaneous dual-writes to Oracle and Cassandra to achieve Zero Downtime Data Migration.

RhinoSource then built the DevOps scripting and deployed the new Cloud architecture on Amazon AWS across development, load and stress testing, and production environments. Comparative load and stress testing on AWS demonstrated that the Cassandra-backed platform outperformed the legacy Oracle architecture while delivering linear scalability and active-active data centers at a lower price point.

We continue to support application quality assurance and performance tuning as new features land and as database upgrades are released and applied at customer sites.