API Streaming & Sustainable Growth

Describing environmental impact of API streaming based on benchmarks, or how to use eco-design principles in a deeptech, delivered in Apidays.io Paris Jan 2018 We discussed it with Mike Amundsen in this panel :https://youtu.be/zsZEpCSTojg

Transcript

1. COVER Streaming API for sustainablegrowth ApidaysParis, Jan 2018
2. The API Economy Use APIs to create apps Developers Providing access to the business assets APIs Consume data through web APIs Apps Increase revenue by using apps with business assets End Users 1 API in 2005 / 18.000 in 2017 Only 10% of all APIs (90% private) ProgrammableWe b Google: 5 B/day (2015) Xignite: 5 B/day (2017) API Calls Salesforce: 50% of $8.4B Expedia: 90% of $8.7B API Revenue Use APIs to create apps Developers Exposable enterprise services Business Assets "By 2020, 50% of organizations managing APIs will incorporate event-driven APIs in their architecture. » Gartner, 2017
3. static UX real time U.X. Enhanced UI for applications
4. The Challenge New Data Time New Data How to take the best out of the API economy when your data is time- sensitive? USELESS API CALLS 98% of API calls retrieve data the client already has SERVER CPU WASTE Up to 90% of the CPU power is used to perform useless tasks Server Client 1. data request 2. response Server Client API data is retrieved through polling. Polling doesn’t work in real time 1. To try to stick to reality,clients constantly hit the API for new data. It’s the Polling Madness. 2. Studies show this pattern is inefficient and waste server and client resources 3.
5. The Solution A scalable and secure proxy turning APIs into streams of real-time data. Clients no longer have to poll the API. Streamdata.io takes care of that and only pushes new data to the client as soon as it’ s available. API providers no longer need to build an infrastructure to handle pollings. Streamdata.io polls the API and distribute data to millions of clients
6. Machine Learning Platform Predictive Analysis Event-driven A.I. Build streaming models with events Social networks feeds Market data Alternative data "By 2020, 50% of organizations managing APIs will incorporate event-driven APIs in their architecture. » Gartner, 2017
7. Efficiency Monthly Costs without with API servers $4,000 $400 API network $3,000 $300 Streamdata.io $0 $3,000 Total API Cost $7,000 $3,700 Case study of investment app with 1B. API hits/month Case study of building and running a streaming backbone $0,00 $0,25 $0,50 $0,75 $1,00 $1,25 $1,50 $1,75 year 0 year 1 year 2 year 3 Millions Without Streamdata.io With Streamdata.io $1.400.000 for 24 months running $180.000 for 36 months running -87%
8. 8 Data vendors Provide a streaming version of your API to your most demanding customers Streaming per month $999 Starter Medium-sized Large $x $xx $xxx per month per month per month Our great API Discover our plans Activity on API Revenue Increased revenue Reduced Cost Increased user satisfaction (no limit, real-time, push) Starter Medium-sized Large $x $xx $xxx per month per month per month Streaming per month $XXX Our great API Discover our plans Activity on API Revenue
9. Partners & Clients
10. Our Values: People, Profit, Planet J.Elkington, 1994
11. Our Values: People, Profit, Planet http://streamdata.io/blog/ui-latency-json-patch/
12. People: Real-Time Transparency For an electronic trading platform, being 10ms behind competitors can result in a 10% drop in revenues. With 100ms you might as well shut down TABB Group
13. Our Values: People, Profit, Planet
14. Environmental Impact Functional unit Updating stock price on smartphones during one day. 37,921 daily active users updating 21 stock prices per session. Each session lasts 12 minutes. LCA Model EU Lifecycle assessment: http://www.eebguide.eu/ Greenhouse Gas emissions (GHG) for Climate change (GWP) Water consumption for Abiotic resource depletion (ADP) Energy for Abiotic resource depletion (ADP)
15. Benchmark without
16. Benchmark with
17. Compare Without With
18. Environmental Impact over 1 App GHG 1,443,000 kms in car Water 117,500 showers Energy 327 French houses
19. COVER www.streamdata.io Eric HORESNYI C.E.O. eric.horesnyi@streamdata.io