With NS1 traffic steering capabilities we can do great things, and if, on top of it, we add the integration capabilities between SevOne and NS1 we can do magic!
Let me tell you a story
I'm sure you guys have heard a lot about ML and things like chatGPT and WatsonX lately, these are technologies that will revolutionise the world we live in, and this is typical conversation point when we are visiting customers. However, in one of this meetings, one customer asked me what are we doing with ML in the DNS world? Of course I asked why he was interested and his answered that for his company the performance of their public web applications is key and the fact that they were being reactive rather than proactive was a problem for them. And, of course, with all the noise that ML was making, he was curious if there was a way to utilize this technology today to make their performance even better.
Example of data collected by NS1
Certainly I was very happy with that question, because today there is something that we can do to make their web applications more proactive and to avoid future performance issues, and the answer is using SevOne analytics.
What is SevOne?
IBM SevOne NPM is a timeseries data monitoring solution that can do multiple things (I always say that there are more than a thousand unique features), but for this particular use case:
- Can collect data from anywhere, any source (including NS1 platform)
- Analyses the data using different algorithm (including Machine Learning)
Those two features will allow us to collect all the data that NS1 is collecting, from Real User Monitoring (RUM) data from the different cloud and CDN providers, to query volume information from each of the zones, records and networks available in NS1. Once this data is available in SevOne, we cannot only report and create amazing dashboards with them, but also SevOne will analyse the data out-of-the-box giving us some insights to help us prevent future performance degradations.
Normal Behaviour Analytics
How does it work?
There are four main steps to proactively avoid future performance degradation.
Step 1
Send all the data that you are collecting in NS1 to SevOne. This can be done in different ways, but possible the easiest one is by using the NS1 API to send the data to SevOne's API. We can use scripts for this, or we can use IBM Rapid Network Automation to use the nocode automation platform to get the data from NS1 and send it to SevOne using building blocks. This way we don't need to learn APIs or to do any Python or NodeJS script.
IBM Rapid Network Automation Workflow
Step 2
Once we have the data in SevOne, it's time to analyse this data. This is something that will happen automatically in some algorithms such as 'normal behaviour' analytics or calendar analysis, and in some other situations we will need to create the analytics ourselves, for example on the forecast projections we will need to define some parameters such as lookback time or time to project.
Forecast Projection Example
Step 3
Send the results of our analysis to NS1. Again, there are different ways to do this, but possibly the best way would be using Data Feeds, so we can attach the results of these analytics to metadata.
Data Feeds Example
Step 4
Finally, configure NS1 to steer the traffic based on the data received from the Data Feeds, and out-of-the-box functionality that helps us to steer traffic based on lots of different options (location, cost, status...) including based on data sent by other systems such as SevOne.
Traffic Steering Options
Apply ML Analytics today
All these steps are available today, you only need the tools to implement it and make sure that you go (or steer in this case) around performance problems by analysing the normal behaviour of your network and how they will behave in the near future.
NOTE: If you want to know more, check out this video where we show the integration between SevOne analytics and NS1
https://community.ibm.com/community/user/aiops/viewdocument/sevone-automated-network-observabil-1?CommunityKey=5a69b903-4d5f-4aaa-a21d-018ab33b60e2&tab=librarydocuments
#NS1
#IBMSevOne
#TechnicalBlog