Effortless Serverless 

The journey Current development and deployment Problems Washing machines Servers Serverless 

Random startup company • Education • Innovative • Investments (millions $$$) • High level experts, publications, patents • Switched to new stacks a few times already, each “better” than the previous one • 100,000s users 

Organization Customers (Angular) API Admin (Angular) Company site (Wordpress) Elastic Search “State of the art” Other services Owners (Angular) 

Infrastructure VPS (server rental) • Heroku • EC2 • Azure Environments (per server) • Development • Staging • Production 

Scaling, balancing? • Capacity planning 
 At best random guesses. 100,000 users. 
 Estimations - at most 20,000 are going to login at the same time, horrible mistakes • Environments setup
 dedicated DevOps guys = dedicated cursing machines • Deployment
 ssh, scp everything • Scaling, Balancing
 the hell 

Problems • Taking 15 mins to 1 hour to deploy a “bug” to all production servers • One service eating up the server memory, other dying and not logging • Most of the time just waiting for a request • Recovery time (2-7h) • Configuring a new server took 2h • Pressure / Yelling • Angry phone calls / messages at 4 am - 
 “Are you awake? The production crashed.” 

Servers are most similar to? 

Washing machines What the hell they have in similar with the current development and deployment issues? Quite a lot. 

Washing machine • Processing machine - cleans clothes • Receives various types of clothes • Has a maximum amount • Can mix different types of clothes • You can buy various types from a big one to handle more, to small ones 

Problems • Takes from 15 minutes to 2h to wash all your laundry • One cloth color can spoil the other - your special princess dress • Waiting, until it receives clothes • Can get stuck or broken (then you have to call the plumber)
 Recovery time: > 1 day • Buying a new washing machine and setting it up >2h • You can buy a big one to handle more, but then its dumb to turn it on just for one shirt f.e. - so you usually wait • Your partner yelling at you for breaking it 
 (definitely not from a personal experience) 

Looks familiar? 

• Taking 15 mins to 1 hour to deploy a “bug” to all production servers • One service eating up the server memory, other dying and not logging • Most of the time just waiting for a request • Recovery time (2-7h) • Configuring a new server took > 1h • Hexa-core for an Angular app? Seriously? • Pressure / Yelling • Angry phone calls / messages at 4 am 
 (“Are you awake? The production crashed.) • Takes from 15 minutes to 2h to wash all your laundry • One cloth color can spoil the other - your special princess dress • Waiting, until it receives clothes • Can get stuck or broken (then you have to call the plumber)
 Recovery time: > 1 day • Buying a new washing machine and setting it up > 1h • You can buy a big one to handle more, but then its dumb to turn it on just for one shirt f.e. - so you usually wait • Your partner yelling at you for breaking it 
 (definitely not from a personal experience) Server issues Washing machine issues Can’t be the same… 

The rise of the laundry
 ”The evolution of the laundry industry” First “Washateria” 1934. Mass laundromats 1950s First “Fluff and Fold” services 1970s Mass “FnF” and the decline of laundromats 2000s 

And servers? First “VPS’ 2001. The “Cloud” 2005 Serverless? ? 

So serverless is? • The new “Fluff and Fold”? • The new “cloud”? • Silver bullet? • Holy grail? • Cure for cancer? 

This? 

Serverless is a method of deploying and running applications on cloud infrastructure, on a pay-per-use basis and without renting or buying servers. Serverless But how does not having servers help? 

Serverless does not exclude the existence of servers. It just abstracts away the need to physically rent or buy a server, including setting up its OS, environment, installing services, and so on. Serverless 

How does that work? Containers 
 Isolated environment that contains everything your application needs to run - as plant pots. 

Pay per use? 

Average API server usage Utilization Intervals 35% 60% 80% 06 09 12 15 18 70% 45% 21 00 03 06 0% 100% Looks good enough? Price: 100$ 

What if we change the perspective? 

Average API server wasted time Intervals 65% 40% 20% 99% 30% 55% 95% 99% 0% 100% Bad enough? Waste 06 09 12 15 18 21 00 03 06 

Wow, just an average of ~ 62.8% wasted resources 

API server utilization + an unexpected peak Utilization Intervals 35% 60% 110% 70% 45% 0% 100% So having new customers Is a problem ?! “Unexpected” peak = new required capacity 06 09 12 15 18 21 00 03 06 

Burn more! Because, why not? 

The “new” API server usage Utilization Intervals 28% 54% 80% 63% 38% 0% 100% We are “safe” now! Buffer New Price: 150$ - (if you’re lucky) 06 09 12 15 18 21 00 03 06 

Need more servers? 

Just do it! We have a million dollar investment! 

The serverless way • Pay per request • No request - no charge • No request - no active server • “Hosting” your app is free 

Providers (most notable) • Amazon - AWS Lambda • Microsoft - Azure Functions • Google - Google Cloud Functions • IBM - BlueWhisk 

Example (AWS Lambda + Claudia.js) 

Other Benefits? 

Refactoring • Makes sense as it has a financial incentive • Better code, less cost • No more statements: “Well, its working good enough already” 

Code efficiency • Slower code has bigger repercussions. • Algorithms having a stronger impact 
 (start reading) 

No content

Thanks!