Discovery in the Age of Network Automation
The facts speak for themselves. Ask any CSP operations team to provide an up-to-date map of all the devices and services running in their network and they will be pained to give a conclusive reply. Discussions with some industry experts indicate that as much as 60% of today’s network resources inventory is deficient. And this statistic does not fully take into account the massive impact of hybrid networks as CSPs accelerate adoption of new technologies such as virtualization and disaggregation in their fixed line networks and future 5G infrastructure.
The operational and commercial advantages to deploying a 100% accurate vendor-agnostic automated discovery and inventory management system are enormous:
- Unify inventory and topology
- Reduce time to deploy infrastructure
- Reduce time detecting/replacing faulty network element
- Enable instant service delivery
- Discover stranded resources and reduce 3rd party resources costs
- Automate network services fulfillment by supporting business and network policies with real-time network data
Nevertheless, there are significant challenges to realizing these benefits. Most inventory systems today must be updated manually. They are not events-driven and as a result are out-of-sync with the network. Significant network information is missing. In many cases data accuracy of legacy inventory systems may be less than 50%, making it difficult to use existing data.
Therefore, it is imperative for today’s CSP to automate the discovery and inventory management process. What are the key functions required for such a system?
- Discover physical and logical network resources — One-click inventory view of all network assets
- All the discovered network data is available for northbound systems
- Sophisticated “check-for-changes” and “delta-syncing” mechanisms to support hyper scale networks
- Discover end-to-end “brownfield” service topology
- Discovered data can be used for a rule-based network-activation in real-time
- Auto-discovery and visualization of Layer 1, 2 and 3 network topology
- Vendor-agnostic for maximum flexibility
Not all discovery tools, however, are equal. Many consider only the physical and logical level. They might also rely on existing and not so accurate inventory systems and files. Even more of an impediment is that these tools require code-based drivers or adapters taking many months to build. Last, these tools are highly dependent on vendor cooperation, reducing flexibility and limiting the pace of creating a comprehensive library of network elements.
Optimally, the discovery automation tool should also enable service level discovery. It should be equipped with intelligent, model driven 100% accurate network discovery capabilities. Time to deployment should be reduced to a few weeks through modeling using a visualized design portal. And instead of dependency on vendor availability, planning, design and even execution tasks should be able to be outsourced to a trained system integrator or even the CSP.
Correct implementation of a 100% accurate vendor-agnostic automated discovery and inventory management system will enable CSPs to overcome many of today’s business and transformational challenges and achieve complete control over their network, leveraging existing infrastructure capital investment for future growth.
For more details , please contact nir.cohen@atrinet.com.