The Inspiration Behind Failing – Talari’s Humble Beginning

I have been asked many times on what was the inspiration for our solution.  It is a tricky problem to provide high service levels over unreliable service providers.  In all honestly, there were several factors that influenced me in a technical and human way.  I thought I would share one of the human aspects with you now.  I will share more inspirations in later blog posts.

In 2001 I worked as a Director at an international chip company.   At the time we had development centers in the North East, Canada, North Carolina, UK, Northern and Southern California working collaboratively designing new silicon/software solutions for network equipment vendors.  I was deeply involved in the founding and staffing of our newest Bangalore office and was spending quite a bit of time on the project.

As is often the case, getting good quality private networks to the Bangalore office was problematic and delayed.  As a result out of necessity, we had to make due with a site to site IPSec VPN.  Although not ideal, the IPSec VPN was good enough for us to bootstrap the office.  An IT administrator in the team, Ron (name is changed to protect Ron), was very enthusiastic on the possibilities of expanding the use of IPSec VPNs for all our other domestic and international offices.  He and I became friends while working on the Bangalore build out and we talked about the possibility and the potential.  I was intrigued.

This was, after all, after the tech bust of 2000 and the company was constantly looking to save money in any way possible.  So Ron’s position on removing the expensing ATM/MPLS private network and replacing it with appliance site to site IPSec VPNs was well received by me and the executives, since it had the potential to positively impact the business and maybe even help us prevent job losses.  So the project commenced and the enterprise started to replace the private network with IPSec VPNS.

To say it simply, the project was a disaster.  The IPSec VPN underperformed dramatically and the productivity and ability for site to site collaboration suffered. Ron and the company separated ways soon after.   My friend’s failure affected me and got to me. This is the source of my common saying “friends don’t let friends drive on single link IPSec VPNs.”

That said, as is my tendency around such things, this failure got me thinking.  Why did the project fail so miserably?  The bandwidth was unreliable, packet loss was intermittent, links were oversubscribed, there was no concept of QoS, and the IT team had no insight as to how to solve issues when they occurred.  The broadband service providers had no accountability for bad networks.

It made me think…  what if we could add an intelligent layer between the users and utilize multiple service providers at each location?  What if we could provide first and last mile bandwidth reservation to prevent oversubscription on the multiple diverse links? What if it would encrypt just like IPSec, but in addition would monitor all the traffic for application and network behavior?  What if then we could use that real time information to broker the traffic to the best capable service provider instantly? What if we could detect packet loss and retransmit traffic on alternative networks when a network is temporarily lousy?  What if for key applications we could redundantly transmit traffic and squelch duplication on the far side of the network? What if we could measure the one way latency and then adjust time sensitive applications away from high latency providers but allow latency tolerant bulk traffic to still use it? Lots of what ifs started swirling around in my mind on this very failed IT project and how it impacted the company and Ron’s career.

When I travel around the world and I share this story with IT administration teams I quite often see in their eyes that they too had a Ron in their worlds who leapt too far to IPSec VPNs and failed.  We also look at each other and know that those failures were inspirations as to why we are talking now about our proven solution that achieves and fulfills Ron’s vision of a cost efficient and hyper reliable, secure and intelligent network for the users and employers.

Today the networks are better and more available. Our technique and solution is proven and easy to take for granted.  That said, if not for pioneers like Ron, who ventured using a prior much more primitive generation of technology and failed, Talari would never have happened.

Categories: Internet as WAN (MPLS Alternatives), Intelligent WAN

Tags: , , ,