I was working for decades in the mobile phone and general embedded markets where I
honed my skills in producing small form-factor commodity solutions.
I had also previously worked at a company where they were building rack mount 1U units based
on Intel platform ITX motherboards (which were cheap) and using Linux to make a VoIP rack server.
As well as this, I worked for some well known brands including Siemens, BBC and Sony, where I learned about the importance of robustness testing when
providing solutions on which you need the customer to rely for many years after the product is released.
NetLinux is a Linux distribution I created 12th August 2007 to help with providing Linux solutions to my customers.
I also created a company in 14th October 2008 and the online content was served on a private cloud I created called "Virtual Rack".
Virtual Rack was a platform on commodity Intel platform hardware with NetLinux and
open-source VirtualBox, and served using my existing broadband service.
This was providing websites and other online facilities for some local organisations in Reading.
Back then I was using a capped, scheduled broadband service for which I was charged during the peak hours.
I needed to see what was being used over this service, so it was then that I got into network monitoring.
Some open-source command-line sniffing tools (tcpdump and tcptrack) running on a NetLinux based virtual machine showed
all that I needed to work this out.
When I joined a network company in 2013 as a support engineer on their premier customer account, it
seemed they had issues with providing a reliable network service to this financial customer.
It became clear that I could re-use my "broadband monitoring" solution to diagnose issues on their networks.
The problem was that my employers did not have any compute facilities in their network on which you could install any software to do monitoring.
So this is where the cheap ITX board solution I used before would probably work for this.
I found a small, cheap Intel-based ITX motherboard-base box unit which was a "thin client PC" from Jetway.
Add a laptop hard disc and some RAM and I had a hardware platform for my simple broadband monitoring machine.
I set this up with my software which was basically NetLinux and tcpdump on these Jetway boxes.
I was not required or asked to do this, nor was it any part of my job as a support engineer.
I did it to help my employers and their relationship with their customer whose account I was hired to support.
The NAPI Project
Using those years of pre-existing experience in other companies and my own simple broadband monitoring machine I
created a sniffer to solve my employers problem.
It was created because I could see my employers had issues with networks which I had seen elsewhere, so
decided to re-use that for them to help them.
My original simple broadband monitoring machine had never been used in an enterprise environment, so I
needed to refine it for that and get it through some sort of testing plan.
This means it needed to be a stable device which was easy for other people to deploy and use.
A fixed tight specification which didn't aspire to doing too much put through rigorous testing before it can be considered ready.
Although, it was refined and tested successfully, it was prone to dropping packets and the statistics it provided were not very accurate.
It did, however, provide a good enough insight to the problem of drops due to overuse of bandwidth on carrier connections.
It also provided me with enough information to tell the customer what level of bandwidth they needed to fix their problem.
The sniffers were so popular that other people in the company and even my employers' customers were finding out about them.
Since the hardware was branded "Jetway" the sniffers were being called "The Jetways".
This was a problem as then the customer thought they could buy a "Jetway" thinking it was one of my sniffers.
This caused confusion, so I decided to create a name for the project so people had a kind of "brand" with which they could associate the units.
Since I created the project, I decided to have a bit of fun with naming it.
In support you always feel like you are catching the doo-doo, like a babies nappy, so I created an acronym which sounded like nappy.
That name is "Network Application Performance Inspectors" or "NAPI" for short.
NAPI wasn't anything new in any way, in fact Centurylink also had an internal project called "Chariot" which did pretty much the same thing.
Chariot was expensive and complicated to deploy and use and the laptops with Windows were not designed and tested for long term capturing.
Even the creator of the Chariot project was looking to abandon the Laptop+Windows+Wireshark thing and use NAPI instead, due to
it's simplicity, reliability and low cost.
The NAPI project was not a research project into a new technology or product, nor was "a NAPI" something you could ever purchase anywhere.
It was simply an internal joke name I gave to an enterprise refined version of my simple broadband monitoring machine which
I used to help my employer with network issues.
The NAPI project became so popular with my employers' customers that they were asking when NAPI units would be available for them to purchase.
The Birth of MONI
For 2 years I and many other employees attempted to get my employers to support this popular project.
I was expecting them to see that this was an opportunity to solve problems by selling units to their customers.
This would not only create a new revenue stream, but also raise my employers credibility.
A win-win situation.
The response I eventually got from the management was that my employers were not in the business of creating and selling this kind of product.
I realised the opportunity I was giving my employers to benefit from this project was never going to be pursued.
That original solution, created back in 2008 and finally tested in an enterprise network, seemed
to be something which would fill a gap in the market but my employers were not going to be the organisation to do it.
So I decided to leave that company and sell my, now refined, original broadband monitoring product under my existing company NetLinux Ltd.
To bring an enterprise product to the market which is priced so a customer can distribute units everywhere in their network.
So my customers can have intimate knowledge of every packet without costing them millions to own and operate.
That product is MONI.