Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

So you think you can scale containers

174 views

Published on

It's a question with an answer. The answer is yes, if you embrace service mesh.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

So you think you can scale containers

  1. 1. SO YOU THINK YOU CAN S C A L E@F5Networks containers?
  2. 2. ☺   Why do we scale?
  3. 3. =
  4. 4. 100 Milliseconds Slower -1% SALES -0.2% SEARCHES -2% CONVERSION $660M $45M $244M H/T James Urquhart, SOASTA Data: Gartner, Walmart
  5. 5. 1 Minute of Downtime Data: Emerson Power Costs an average of $7300 Average total cost of downtime per year across industries PRODUCTIVITY IT PRODUCTIVITY LOST REVENUE $53,608$140,543 $183,724
  6. 6. UP OUT How do we scale today?
  7. 7. But architectures and apps are changing THEN MONOLITHIC MICROSERVICES & APIs NOW
  8. 8. And so are environments THEN NOW DATA CENTERS & SERVERS CLOUDS & CONTAINERS
  9. 9. COMMS INCREASES EXPONENTIALLY THEN NOW
  10. 10. SMALLER CAPACITY MEANS MORE CLONES THEN NOW
  11. 11. RAPID CHANGES MAKE SCALING HARD NOWTWO MINUTES AGO
  12. 12. REVERSE PROXY THEN NOW
  13. 13. FORWARD PROXY • NATIVE CONSTRUCT • BASED ON IP TABLES • LIMITED to IP:PORT
  14. 14. REVERSE PROXY • EPHEMERAL CONSTRUCT • BASED ON IP TABLES • LIMITED to IP:PORT
  15. 15. SIDECAR PROXY • PLUG-IN CONSTRUCT • FORWARD PROXY • L7 CAPABLE
  16. 16. LAYER 1 LAYER 2 LAYER 3 (IP) LAYER 4 (TCP) LAYER 5 LAYER 6 LAYER 7 (HTTP) LAYER 7.5 (TAGS) LAYER 4.5 (MESH) LAYER 2.5 (OVERLAY - VXLAN) TRADITIONAL OSI STACK Assumes underlying protocols behave consistently and reliably. Relies on proper adherence to and reliability of lower order protocols. CONTAINER OVERLAY STACK Assumes underlying protocols exist but are unreliable. Implements its own routing and reliability constructs to avoid failure of and reliance on lower order protocols. BUILDING A SERVICE MESH
  17. 17. A SERVICE MESH Circuit breakers Auto-Retries A network of sidecar proxies that form a reliable method of scale that includes: Health monitoring HTTP Routing
  18. 18. HTTP ROUTING Like Google maps, but for your apps URI & method host custom Cookie GET /path/to/my/service/with/variables http://host.domain.xxx request headers X-API-version: 1 Cookie: 09124890tlsgkq3i3KJSDLFJ)*@$#() You encrypt your cookies, right? RIGHT?!
  19. 19. RETRIES DO OVER! • CONGESTION • CAPACITY • FAILURE • REJECTION ☺ 
  20. 20. CIRCUIT BREAKERS STOP TRYING ALREADY
  21. 21. MONITORING IF I’M NOT HOME, DO YOU STILL DRIVE TO MY HOUSE? Service is unavailable Service is slow Service is returning errors
  22. 22. SCALING CONTAINERS SMARTER and FASTER : SERVICE MESH Must be L7 (HTTP) fluent Delegate instead of dictate Monitoring is not optional Collaborative not constant
  23. 23. THANK YOU! @F5Networks

×