This week, our Chief Strategy Officer and Group President of AT&T Technology and Operations, John Donovan, introduced ECOMP. And then, we delivered a whitepaper to the industry on the work.  So, what is ECOMP? As our whitepaper details, ECOMP lets us rapidly on-board new services (created by AT&T or third parties) that our customers want. It provides a framework for real-time, policy-driven software automation of network management functions.  

There is no shortage of network cloud management visions / partial solutions / open source initiatives out there in the industry. So why did we create our own?  Here’s why:

  1. Speed and flexibility.  We want to be able to quickly onboard, set-up and run VNFs.  We want to be able to do it at scale, and we don’t want to run several different solutions for each of the various types of VNFs that are required in our network.
  2. The existing solutions or initiatives are based on proprietary technology or are only partial solutions to the real needs.  A holistic solution must include orchestration, control (application, infrastructure, network, and storage), management (service, fault, and performance), and policy (coding our unique expertise into the running of the network).
  3. Any open source effort needs a strong community and a credible third party to support it at scale.  Building a community is a lot of work. Even with a vibrant community, there needs to be an entity that takes on the support required across different service providers: enhancements, updates, troubleshooting, etc.
  4. Solutions must be vendor and VNF neutral.  That means we need modelling language for VNFs, the way the data models are described, the way that the measurement details are delivered, etc. are open and common, and are not specific to any vendor or class of VNF.
  5. Solutions must address service provider needs and operate at the scale required by these service providers.  The old model of buying a product from a vendor and then spending additional money for the measurement, operational support systems, management systems, and training your folks (and the vendors sometimes) about to how to run it at scale.  That isn’t cutting it. We need a better, more complete, more automated, more flexible model.

There are 8 major software subsystems in ECOMP.  They cover areas like:

  • orchestration of virtual machines (VMs) for compute, networking, storage, and measurement
  • controllers that implement the network plan and configure and monitor applications; data collection and analytics that monitor KPIs and inform decisions on policy
  • policy, where we embed our intelligence as a scaled network operator help automate certain decisions
  • all the data for the cloud infrastructure and the VNFs is collected in a geo-redundant data base, consisting of active and available inventory. 
  • service design and creation – a design studio to facilitate service and infrastructure design, allowing re-use across the enterprise 

Stay tuned to the blog as I plan to write more about each subsystem to explain what each does and how they relate to one another.

ECOMP is complementary to OpenStack.  OpenStack is only the cloud layer with some progress in Murano for applications, but that is still young.  ECOMP works in a hybrid environment – OpenStack and VMWare, CPE devices, bare metal compute, and physical elements. 

Importantly, as a model-driven platform, this framework costs less than maintaining existing network systems.    Specifically, the design and creation capabilities allow services to be defined with minimal new IT development.  This means less capital expense.  The real-time data collection, analytics, and policy functions provide significant automation of network management needs.  It enables the detection and correction of problems in an automated fashion, contributing to reductions in operating expense. 

Our work on this software is a direct result of our commitment to becoming a software-centric organization. We’re moving from a company that has lots of customized hardware to one whose network infrastructure and services are controlled by flexible software.  We’re offering a glimpse at part of our playbook with this ECOMP white paper. And, we’re asking for feedback on this direction from the larger cloud community.  If you haven’t yet, please download our whitepaper at  Let us know your thoughts at:  Tell us if you want to help drive the innovation needed to meet the demands of our industry’s future. 

Chris Rice
Chris Rice Senior Vice President – AT&T Labs, Domain 2.0 Architecture and Design