Availability Concepts

Question 1

Fault tolerance

Answer 1

• Maintain uptime in the case of a failure
  
  • If a problem occurs, what happens?
  • Can degrade performance
• Fault tolerance adds complexity
  
  • The cost of managing the environment increases
• Single device fault tolerance
  
  • RAID, redundant power supplies, redundant NICs
• Multiple device fault tolerance
  
  • Server farms with load balancing
  • Multiple network paths

Question 2

Load balancing

Answer 2

• Some servers are active, others are on standby

* If an active server fails, the passive server takes its place

Question 3

Redundancy and fault tolerance

Answer 3

• Redundant hardware components
  
  • Multiple devices, load balancing power supplies
• RAID
  
  • Redundant Array of Independent Disks
• Uninterruptible power supplies (UPS)
  
  • Prepare for the disconnections
• Clustering
  
  • A logical collective of servers
• Load balancing
  
  • Shared service load across components

Question 4

High availability

Answer 4

• Redundancy doesn’t always mean always available
  
  • May need to be enabled manually
• HA (high availability)
  
  • Always on, always available

• May include many different components working
together
• Watch for single points of failure

• Higher availability almost always means higher costs
• There’s always another contingency you could add
• Upgraded power, high-quality server components,
etc.

Question 5

NIC teaming

Answer 5

• Load Balancing / Fail Over (LBFO)
  
  • Aggregate bandwidth, redundant paths
  • Becomes more important in the virtual world
• Multiple network adapters
  
  • Looks like a single adapter
  • Integrate with switches
• NICs talk to each other
  
  • Usually multicast instead of broadcast
  • Fails over when a NIC doesn’t respond