AWS FAULT TOLERANCE & HIGH AVAILABILITY

1 S. FAULT TOLERANCE . Amazon Web Services provides services and infrastructure to Amazon EC2 provides infrastructure building blocks that, by W. A ce build reliable, FAULT -tolerant, and highly available systems in the qualities have been designed into our services both themselves, may not be FAULT -tolerant. Hard drives may fail, power supplies may fail, and racks may fail. It is important to use enes by handling such aspects without any special action by you and by combinations of the features presented in this document to f Reectur e r & high AVAILABILITY providing features that must be used explicitly and correctly. achieve FAULT TOLERANCE and high AVAILABILITY . h it 2. Arc C. E. n zo FAULT TOLERANCE and high AVAILABILITY of a S.

2 M B. Amazon Web Services A. E. A. S3. m n a zo zo g n a n Most of the higher-level services, such as Amazon Simple ci m an A. Storage Service (S3), Amazon SimpleDB, Amazon Simple al B. Queue Service (SQS), and Amazon Elastic Load Balancing ad 3. (ELB), have been built with FAULT TOLERANCE and high o S. L. n ic AVAILABILITY in mind. Services that provide basic infrastructure, zo st la such as Amazon Elastic Compute Cloud (EC2) and Amazon a E. m Elastic Block Store (EBS), provide specific features, such as A. AVAILABILITY zones, elastic IP addresses, and snapshots, that a FAULT -tolerant and highly available system must take 2 advantage of and use correctly. Just moving a system into the cloud doesn't make it FAULT -tolerant or highly available.

3 A. DaRep ne y ta lica A. Zo lit La te ye d EB. bi A. r ts Ab m ila S. ho a zo ilit ps n yt EC. va A. na m o S. 2. a fa 4. zo A. il o se n ve a tab r ddeeunn A. ppee ec vo Darver Ap pl nnd es id Se ica 2. ion tio s deenncsar (s n 4. cat ta O. ies y cie n 2 R n d i b l re ep Fa y). p 2 -a lac ilu Aprver tta e r B. ch in e: b Se Ap pl ica vvoo sta lu lum n mee cea 1. Werver tio annd d B n (a Se ct iv ne y e ). Zo lit BaElas bi ila En lan tic Ap va d cin Lo El pl ica (R. En Us g ad A. as e d tio p er la Us tic n ce 3. m er e IP. n t). Load balancing is an effective way to increase the system, you will need a good understanding of zone ability to re-create the data. Amazon Elastic Block Store System 1 AVAILABILITY of a system.

4 Instances that fail can be replaced seamlessly behind the load balancer while other dependencies. (EBS) offers persistent off-instance storage volumes that are about an order of magnitude more durable than on-instance Overview instances continue to operate. Elastic Load Balancing can be used to balance across instances in multiple AVAILABILITY 3 Elastic IP addresses are public IP addresses that can be programmatically mapped between instances within storage. EBS volumes are automatically replicated within a single AVAILABILITY zone. To increase durability further, point-in- zones of a region. a region. They are associated with the AWS account and not time snapshots can be created to store data on volumes in with a specific instance or lifetime of an instance.

5 Amazon S3, which is then replicated to multiple AZs. While AVAILABILITY zones (AZs) are distinct geographical Elastic IP addresses can be used to work around host or EBS volumes are tied to a specific AZ, snapshots are tied to 2 locations that are engineered to be insulated from AVAILABILITY zone failures by quickly remapping the address to the region. Using a snapshot, you can create new EBS. failures in other AZs. By placing Amazon EC2 instances in another running instance or a replacement instance that was volumes in any of the AZs of the same region. This is an multiple AZs, an application can be protected from failure at a just started. Reserved instances can help guarantee that effective way to deal with disk failures or other host-level single location.

6 It is important to run independent application such capacity is available in another zone. issues, as well as with problems affecting an AZ. Snapshots stacks in more than one AZ, either in the same region or in are incremental, so it is advisable to hold on to recent another region, so that if one zone fails, the application in the Valuable data should never be stored only on instance other zone can continue to run. When you design such a 4 storage without proper backups, replication, or the snapshots.