Telematics Container Cloud Platform (TCCP)

Infrastructure

Verizon Connect · Mar 2016 – Oct 2018

Multi-site Active-Active containerized internal cloud platform with eBGP Clos architecture, achieving 100% uptime over 1.5+ years and reducing deploy cycles from 30 days to 8 hours.

Overview

Designed and built TCCP, a multi-site Active-Active containerized internal cloud platform at Verizon Connect (Telematics). Led the creation of the first DevOps team in Telematics, hiring, training, coaching, and mentoring 8+ employees and contractors. The platform was designed for high resilience and fault tolerance, capable of handling multiple concurrent hardware/software failures within and across sites. It leveraged eBGP Clos network architecture, Consul DNS-based service discovery with multi-site failover, Docker Swarm container orchestration, and Ceph distributed storage. The entire infrastructure provisioning and configuration was fully automated end-to-end using Ansible. Represented Verizon Connect on the OpenSwitch Project Technical Steering Committee under the Linux Foundation and presented at Open Networking Summit North America 2018.

Key Highlights

Multi-site Active-Active containerized internal cloud platform
Highly resilient/fault-tolerant infrastructure handling multiple concurrent hardware/software failures
eBGP Clos network architecture with Consul DNS-based service discovery and multi-site failover
Docker Swarm orchestration with Ceph distributed block/file/object storage
Fully automated E2E network and server infrastructure provisioning using Ansible
100% uptime over 1.5+ years in production
Deployment cycle reduced from 30 days to 8 hours
Represented Verizon Connect on OpenSwitch Project TSC under the Linux Foundation
Presented at Open Networking Summit NA 2018
Led creation of first DevOps team in Telematics (8+ members)