An essential part of IT architecture consulting is familiarization with documentation. It is probably no news to anyone that documentation in the field is quite varied with quality, content, utility, and timeliness.

As an example, we were planning to modernize a large factory area data network. The documentation stated that the new fiber network was built to increase capacity at the mill site and staff was proud of it. As I watched the documentation, the idea arose to check. By afternoon, it had become clear to the project team that data was not moving in the invested fiber network. All traffic was further routed through the operator’s MPLS network, where additional capacity had also been invested. Similar situations occur more and more frequently.

Manual documentation and processes have come to an end

The most common reasons for this are easy to list: Process and development staff turnover, constant high workload, too few expert staff, multiple simultaneous and partial network change projects in a multi-vendor environment, fast failover, e.g. static routing to quickly fix issues comes permanent, documentation available to all stakeholders, multiple sub-vendors, and lack of overall view in network management.

Even industry agree documentation is vital for network investment, capacity management, rapid troubleshooting, efficient automation, and compliance it cannot afford it.  The Nordic wage level is too high to solve the old operating model. Also important to note that future enterprise networks are much more dynamic and, together with cloud services, more complex. For example, SD-WAN utilizes multiple access networks simultaneously and routes data according to its own decisions, as the case may be. Adding to this the rapid change in business, manual documentation and separate control processes have come to an end.

 

Network automation and artificial intelligence as a solution 

As a consultant, I have the tools to determine the true state of your network but in large networks with thousands or tens of thousands of devices need network automation platform is required.  For example, the NetBrain Technologies automation system builds its operating model by analyzing about 12,000 variables per network device. Thus, a network of 1,000 devices already has 12 million variables, all of which are constantly monitored by NetBrain Artificial Intelligence (AI) and Machine Learning (ML). A key approach is to create a web automation process that everyone in the organization can utilize. The benefits are significant:

  1. Hybrid (IP, SDN, Data Center, Cloud) networks and documentation are always automatically up-to-date for Layer 2 and Layer 3  maps with relevant details. There are no human errors in the documentation.
  2. Various network maps can be easily utilized widely in different functions of an IT organization at the same time. Everyone has the same snapshot.
  3. Clear network maps and visual views already enable first-level Service Desk staff to act as second-level experts. Individual device manufacturer tools or CLI skills are not required when using the NetBrain interface. The whole network is easy to see and understand.
  4. ITSM integration (eg ServiceNow, Remedy) and NetBrain artificial intelligence allow for the automatic creation of a fault ticket. The ticket follows links (URLs) to the network documentation and to the network issue. The goal is to manually reduce the number of failures created by min. 90%. Also, CMDB can be synchronized.
  5. Third-level network experts will immediately get a more detailed description of a potential problem or need for change. With NetBrain automation tools (Runbook), third-level experts can quickly automate tasks without programming, scripts or standalone software robots.
  6. Network growth is well under the control of experts. At the same time, the operation of automation is well documented and visible to everyone, keeping abreast of network changes and documentation.

I have seen, evaluate and used hundreds of network management systems. This is the first to work also in practice for NOC and SOC staff, customer service, and IT infrastructure managers and network designers.

In this example, artificial intelligence and machine learning have been harnessed in a brilliant way for IT staff to be a slave without the need to think about moral issues.

Hannu Rokka

5Feet Networks Oy