Availability

AVAILABILITY

efficiency vs cost

presented by Camilo Villegas

AVAILABILITY

Chapter from:Release It!second edition by Michael T. Nygard

If we ask a child how much ice cream they want, the probably would say "Mountains of it!".The problem here is that the child is not seeing the whole image:

• how much does mountains of ice cream cost?
• how bad for his health would it be?

Gathering & documenting availability

Gathering &

Always define the

• Availability 98% 99.99%
• Downtime(min) 864 4
• Downtime $ $21,600 $108
• Added Cost $0 $98,700
• Net Savings $0 $1,289,520

98%(A) vs 99.99%(B) - Availability

WHEN POSSIBLE spending $98,700 to save $1,289,520 seems like a sound financial choice...

SLA

Service Level Agreementdefined commitment between a service provider and a client.quality - availability - responsabilities

also...

System

The totality of the project defined in the SLA

Availability

How the monitoring of availability will be executedAutomatization of the monitoring by emulating the real user.(Like the checklist automatization)

Load Balancing

Improvement to the distribution of workloads, optimizing:

• Resource use
• Maximize throughput
• Minimize response time

Load Balancing

• DNS round Robin
• Reverse Proxy
• Hardware load balancer
• Clustering

Techniques

write a subtitle

DNS Round Robin

Associates several IP addresses with the service name.When a client sends a request to a web page, the DNS will grant him a single IP from a wide list of IPs

Load Balancing

Reverse Proxy

The opposite of a forward proxy.Is a type of proxy server that retrieves resources on behalf of a client from one or more servers. These resources are then returned to the client as if they originated from the Web server itself.

Hardware load balancer

Is a physical appliance used to distribute web traffic across multiple network servers. Routing is either randomized (e.g., round-robin), or based on such factors as available server connections, server processing power, and resource utilization.

Clustering

Is a set of loosely or tightly connected computers that work together.In many respects, they can be viewed as a single system, so if a computer presents a failover, there will be other one to replace it, each node set to perform the same task, controlled and scheduled by software.

Different from load balancing, because in Clustering, each computer is somehow comunicated with the others, in load balancing not.

Thank you!

Any questions?