O slideshow foi denunciado.
Utilizamos seu perfil e dados de atividades no LinkedIn para personalizar e exibir anúncios mais relevantes. Altere suas preferências de anúncios quando desejar.
Protect your app from OutagesRon Zavner, Applications Architect at Gigaspaces                                             ...
AGENDA   AWS and outages   Outage impact   Disaster Recovery – it’s all about redundancy!   Cloudify as a solution for...
AWS USAGE                •   AWS – around 0.5M servers                •   Facebook – less than 0.1M servers               ...
THE OUTAGE PROBLEM4
OUTAGE – APRIL 21, 20115          ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
OUTAGE - JUNE 29, 20126          ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
OUTAGE - OCTOBER 22, 20127          ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
OUTAGE - CHRISTMAS EVE 20128         ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
THAT’S WHAT YOU EXPECT?99% - 3.65 days downtime99.9% - 8.76 hours downtime99.99% - 53 minutes downtime99.999% - 5.26 minut...
OUTAGE IMPACT – DESIGN FOR FAILURESOutage could cost…$89K per hour for Amadeus$225K per hour for PayPal!  10              ...
DISASTER RECOVERY11
MULTI CLOUD12
PREPARE FOR DISASTER RECOVERY•Dedicated expert for DR architecture•Define target recovery time & point•Assume every tier c...
CHAOS MONKEY14
15
CLONE YOUR ENVIORMENT16
CLONE YOUR DATA17
18
CLOUDIFY POSITIONING IN THE CLOUD STACK     Productivity                                                 High productivity...
CLONE YOUR ENV - HOW DOES IT WORK?
EXTENSIVE PLATFORM SUPPORT21        ® Copyright 2012 GigaSpaces. All Rights Reserved
USE ANY CLOUD22
GETTING COMPUTE RESOURCES IN A PORTABLE WAY                                                                               ...
DATA REPLICATION• Cloudify Replicated MySQL Recipe• Generic replication service using WAN Gateway24                ® Copyr...
GENERIC REPLICATION SERVICE OVER WAN                                     London                       New York            ...
26
VERIFI (CURRENT) DEPLOYMENT ARCHITECTURE                                                 PostgresSQL                mod_cl...
TARGET ARCHITECTURE Bootstrap two EC2 clouds in different regions, install the “verifi” application on each. The second cl...
FAILOVER SCENARIO           Cloud #1                                                                         Upon initial ...
DEMO ON EC2 - 5 MINUTES SETUP/* Credentials - You must enter your * cloud provider account credentials */user="ENTER_USER_...
SUMMARY AWS and outages Outage impact Disaster Recovery – it’s all about redundancy!      Cloning your environment – a...
QUESTIONS & ANSWERS    Thank You!RonZ@gigaspaces.com32        ® Copyright 2013 GigaSpaces Ltd. All Rights Reserved
Próximos SlideShares
Carregando em…5
×

19th February 2013, AWS User Group UK, Meetup #3, Managing your apps on AWS: Real life lessons with GigaSpaces, Ron Zanver, GigaSpaces

400 visualizações

Publicada em

Agenda entry: Managing your apps on AWS: Real life lessons with GigaSpaces, Ron Zanver. We’ve all learned Murphy’s inevitable law the hard way – if it can go wrong, it often will! But that doesn’t mean we can’t be ready for such scenarios in the cloud. In this talk, GigaSpaces will focus on the AWS environment, which is dynamic and volatile by nature, and how to maximise your utilisation and minimise downtime. This session will show you how you can architect your cloud-hosted systems to sustain such outages, delving into how to choose the right PaaS for the job, addressing data centre failures, how to avoid single points of failure, and more.

Organiser's commentary: Ron Zanver from GigaSpaces came to talk about the inherent instability of life in the cloud, and what you can do to protect yourself - it's all about good design and architecture. He also introduced us to GigaSaces' new Cloudify product, for abstracting estate management across multiple clouds and cloud vendors.

Publicada em: Tecnologia
  • Seja o primeiro a comentar

  • Seja a primeira pessoa a gostar disto

19th February 2013, AWS User Group UK, Meetup #3, Managing your apps on AWS: Real life lessons with GigaSpaces, Ron Zanver, GigaSpaces

  1. 1. Protect your app from OutagesRon Zavner, Applications Architect at Gigaspaces February 2013
  2. 2. AGENDA AWS and outages Outage impact Disaster Recovery – it’s all about redundancy! Cloudify as a solution for redundancy Demo with Cloudify on EC22 ® Copyright 2013 GigaSpaces Ltd. All Rights Reserved
  3. 3. AWS USAGE • AWS – around 0.5M servers • Facebook – less than 0.1M servers • Google – around 1M servers3
  4. 4. THE OUTAGE PROBLEM4
  5. 5. OUTAGE – APRIL 21, 20115 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  6. 6. OUTAGE - JUNE 29, 20126 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  7. 7. OUTAGE - OCTOBER 22, 20127 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  8. 8. OUTAGE - CHRISTMAS EVE 20128 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  9. 9. THAT’S WHAT YOU EXPECT?99% - 3.65 days downtime99.9% - 8.76 hours downtime99.99% - 53 minutes downtime99.999% - 5.26 minutes downtime 9
  10. 10. OUTAGE IMPACT – DESIGN FOR FAILURESOutage could cost…$89K per hour for Amadeus$225K per hour for PayPal! 10 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  11. 11. DISASTER RECOVERY11
  12. 12. MULTI CLOUD12
  13. 13. PREPARE FOR DISASTER RECOVERY•Dedicated expert for DR architecture•Define target recovery time & point•Assume every tier can fail•Use monitoring and alerts•Document your operational processes 13
  14. 14. CHAOS MONKEY14
  15. 15. 15
  16. 16. CLONE YOUR ENVIORMENT16
  17. 17. CLONE YOUR DATA17
  18. 18. 18
  19. 19. CLOUDIFY POSITIONING IN THE CLOUD STACK Productivity High productivity with full control Heroku PaaS CloudFoundry GAE OpenShift RightscaleDevOps(Automation) Enstratus Puppet Chef IaaS Public clouds (AWS, Rackspace,..) Private clouds (Vmware, OpenStack..) Control 19
  20. 20. CLONE YOUR ENV - HOW DOES IT WORK?
  21. 21. EXTENSIVE PLATFORM SUPPORT21 ® Copyright 2012 GigaSpaces. All Rights Reserved
  22. 22. USE ANY CLOUD22
  23. 23. GETTING COMPUTE RESOURCES IN A PORTABLE WAY compute { template "SMALL_LINUX" }SMALL_LINUX : template{ SMALL_LINUX : template imageId "1234" imageId "us-east-1/ami-76f0061f“ machineMemoryMB 3200 remoteDirectory "/home/ec2-user/gs-files“ hardwareId "103" machineMemoryMB 1600 remoteDirectory "/root/gs-files" hardwareId "m1.small" localDirectory "upload" locationId "us-east-1" keyFile "gigaPGHP.pem" localDirectory "upload" options ([ keyFile "myKeyFile.pem" "openstack.securityGroup" : "default", "openstack.keyPair" : "gigaPGHP" options ([ ]) "securityGroups" : ["default"]as privileged true String[],} "keyPair" : "myKeyFile" ]) overrides (["jclouds.ec2.ami-query":"", "jclouds.ec2.cc-ami-query":""]) privileged true } 23 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  24. 24. DATA REPLICATION• Cloudify Replicated MySQL Recipe• Generic replication service using WAN Gateway24 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  25. 25. GENERIC REPLICATION SERVICE OVER WAN London New York Hong KongIn-Memory Speed Scalable and Efficient High Availability and Self-Healing
  26. 26. 26
  27. 27. VERIFI (CURRENT) DEPLOYMENT ARCHITECTURE PostgresSQL mod_cluster Data Volume Internet EC2 Instance EC2 Instance Cassandra JBoss 4 recipes Data Volume EC2 Instance EC2 Instance Availability region (US-West: Oregon)27
  28. 28. TARGET ARCHITECTURE Bootstrap two EC2 clouds in different regions, install the “verifi” application on each. The second cloud will have a slightly modified (extended) postgres recipe for acting as a slave + no running app servers. Upon the primary zone failure, the second cloud will spin up instances of the app servers and turn the data instance into master, then bootstrapping another “slave” cloud in another zone. replication mod_cluster mod_cluster Data Volume Data Volume Postgres Master EC2 Instance Postgres SlaveInternet EC2 Instance EC2 Instance EC2 Instance Cassandra Cassandra JBoss JBoss Data Volume Data Volume EC2 Instance EC2 Instance EC2 Instance EC2 Instance Availability Region (US-West Oregon) Availability Region (US-East Virginia)
  29. 29. FAILOVER SCENARIO Cloud #1 Upon initial deployment, the primary deployment Cloud #2 of the application will be bootstrapped onto cloud #1, another slightly modified application recipe will be bootstrapped as cloud #2, polling cloud #1 for failure, and acting as a PostgresSQL db slave. PostgresSQL PostgresSQL App Servers Liveness poll Region (US-West Oregon) Region (US-East Virginia) Bootstrap another cloud in a different region using the Region failure same application recipe Turn Postgres slave into used to bootstrap cloud #2 occurs master, Start app server above* instances* Cloud #2 Cloud #3 Cloud #1 Liveness poll App Servers PostgresSQL PostgresSQL Region (US-West California) Region (US-East Virginia )29
  30. 30. DEMO ON EC2 - 5 MINUTES SETUP/* Credentials - You must enter your * cloud provider account credentials */user="ENTER_USER_HERE"apiKey="ENTER_API_KEY_HERE"keyFile="ENTER_KEY_FILE_HERE"keyPair="ENTER_KEY_PAIR_HERE"// Advanced usagehardwareId="m1.small"locationId="us-east-1"linuxImageId="us-east-1/ami-1624987f"ubuntuImageId="us-east-1/ami-82fa58eb" 30 ® Copyright 2012 GigaSpaces Ltd. All Rights Reserved
  31. 31. SUMMARY AWS and outages Outage impact Disaster Recovery – it’s all about redundancy!  Cloning your environment – app stack  Cloning your DB – Replication Cloudify as a solution for Redundancy  Use recipes to work on any cloud  Fast and customized data replication Demo with Cloudify on EC231 ® Copyright 2013 GigaSpaces Ltd. All Rights Reserved
  32. 32. QUESTIONS & ANSWERS Thank You!RonZ@gigaspaces.com32 ® Copyright 2013 GigaSpaces Ltd. All Rights Reserved

×