110 likes | 219 Views
Multi-Server Cloud Environments. How to fight latency and add redundancy. Stop me if this is familiar. <Person a> We need a redundant system. <Person b> The Cloud! I heard that makes it easy! <Person a> Perfect, let’s get IT to make that happen, next order of business...
E N D
Multi-Server Cloud Environments How to fight latency and add redundancy
Stop me if this is familiar <Person a> We need a redundant system. <Person b> The Cloud! I heard that makes it easy! <Person a> Perfect, let’s get IT to make that happen, next order of business... … later that week... <IT> $@%$@# this is hard, why doesn’t the cloud make this easy?!
Complications of Cloud Redundancy • Database Replication and port over • Load Balancing App Servers • Search (SOLR, Sphinx) Scalability and Redundancy • Caching Servers • Shared File Systems
How does the cloud make it easier • Easy to replace damaged “hardware” • Easy to test software changes with temporary servers • Easy to test OS upgrades and switch out • Easy to upgrade “hardware”
How does the cloud make it harder • Hard to predict network proximity • Persistence • Whole Zone Failures
Amazon RDS • Multi-AZ • Point-In-Time Restore • Automatic Replication • Seamless Hardware Upgrades • Automatic Maintenance and Patches
Elastic Load Balancer • Really simple load balancing • Multi-Server redundant • Actually cheaper than a small instance • Server Monitoring built in • Lightning Fast packet level balancing
CloudSearch • Really fast Search • Automatic vertical and horizontal scaling • Basically all core SOLR features • Many plugin SOLR features too • Easy API