Grounded in the Cloud
cancel
Showing results for 
Search instead for 
Did you mean: 

Zen, Cloud Foundry, and the art of designing a better airport security line algorithm

OmriG

Denver_International_Airport_security

As the New York Times reported this week, airport security lines seem to keep getting worse.  Earlier this week, I traveled to San Francisco to visit some Cloud Foundry colleagues, and nearly missed my flight because of a failure mode I had never experienced before when going through the TSA checkpoint.  This checkpoint had two redundant sets of equipment with two different queues behind each. When I was almost at the front of the queue, the conveyor belt for my line stopped working. What happened next was pretty frustrating - even though the other conveyor belt was working and that line continued along fine, our line was completely halted. For 10 minutes! Surprisingly, the TSA staff weren’t trained on how to deal with this, and had no strategy for how to get our line unstuck.

It struck me as a great metaphor for the difference between traditional and cloud-native architectures. What my fellow travelers and I experienced was the ‘traditional’ design – a component breaks down, and the whole system grinds to a halt. Instead, with a ‘cloud native’ TSA line, the breakdown of one conveyor belt is no big deal. Upon realizing that the belt was broken, the TSA agents would immediately route new work (people in line) to the working belt. They would also find ways to reschedule existing work (people who are already stuck behind the broken conveyor belt) to the equipment that was operating. And once they debugged and fixed the broken belt, they would reopen that line.

This is exactly how a cloud-native work scheduling system like Cloud Foundry works. Your micro-services need to be designed for restartability and migration, and if you do that, your app will be built to withstand underlying hardware failures, because the platform knows how to work around these failures.

Fortunately for us, my fellow line-mates and I helped the TSA agents come up with a good merging strategy to get the rest of our line redirected through the other belt, and I was able to make my flight.

This type of illustration helps clarify what we mean when we talk ‘cloud-native’, and reinforces why there’s so much excitement in enterprises and with developers around this movement. And if the TSA started taking a more ‘cloud-native’ approach to designing their systems and processes, we’ll all be happier for it :-)

Join me and the HPE Cloud-Native Platform team at the Cloud Foundry Summit next week in Santa Clara, CA to learn more about how to build cloud-native applications.

HPE is proud to be a Platinum Sponsor for the Cloud Foundry Summit  - Follow @HPE_Cloud for live information during the event and register using CFNA16HPE20 for 20% discount. 

  • HPE Cloud
0 Kudos
About the Author

OmriG

As Vice President of Cloud Products and Services at Hewlett Packard Enterprise, I lead the HPE Helion portfolio, including HPE Helion OpenStack®, HPE Helion Stackato, HPE Helion CloudSystem, and HPE Helion Eucalyptus, extending across our Private Cloud and Managed Cloud routes-to-market. In prior lives I’ve been the General Manager of the Microsoft App Server business; helped start and deliver multiple versions of the .NET and Azure developer platforms; and was responsible for the SQL Server and Xbox platforms as their General Manager. I’ve also worked at three startups, one of which (NEON Systems) IPO’ed in 1999. I live with my wife and three daughters in Redmond, WA, and like rainy winter days because they mean fresh powder on the weekends. Follow me on Twitter @omrig.

Events
28-30 November
Madrid, Spain
Discover 2017 Madrid
Join us for Hewlett Packard Enterprise Discover 2017 Madrid, taking place 28-30 November at the Feria de Madrid Convention Center
Read more
See posts for dates
Online
HPE Webinars - 2017
Find out about this year's live broadcasts and on-demand webinars.
Read more
View all