A Service Level Agreement, or SLA, is a formal document outlining the expectations between a service provider (such as a web host) and a client. The Service Level Agreement is a critical part of any vendor/customer contract. A particular SLA may or may not be contractual (legally binding). In addition to types, levels and quality of service, the SLA specifies which remedies are available in the event service requirements are not met. Publicly viewable SLAs are published by nearly all web hosts and Internet service providers.
A provider's Service Level Agreement outlines the services to be provided, the metrics of measurability for each service, and any penalties and/or remedies should the level of service be insufficient. If a network provider, for example, promises 99.999% uptime of their network, customers should reasonably expect no more than 5.26 minutes of "failure" per year.
In addition to specific, measurable services, the Service Level Agreement will typically include clauses concerning support options, dispute resolution, disaster recovery and so on. The following are some of the services, terms and clauses contained in typical SLAs.
UPTIME/AVAILABILITY. This is a measure of the time a service is properly functioning. The uptime promise may be 99.9%, 99.999% or even 100%. Availability refers to the reliability of a system -- uptime compared to downtime. Different services -- computer servers, cloud storage, network operation -- may have different uptime guarantees.
CREDITS. A credit is any payment or waiver to the customer as compensation for insufficient uptime or some other service failure. In order to obtain a credit, you may need to provide your web host with proof of the down time. There are a lot of different monitoring companies that you can use. Most of them require a paid subscription, but if you need to be monitoring your SLA, it might be worth it. CopperEgg (www.copperegg.com), Manage Engine (www.manageengine.com), and Dotcom Monitor (www.dotcom-monitor.com/sla-management.com/aspx) are just a few of the more well-known options, but there are a lot of other good companies too.
PERFORMANCE METRICS. Such clauses would numerically specify the characteristics of service sub-components such as internal/external network latency and throughput, hard drive I/O rates, storage quotas, CPU speed and availability, maximum simultaneous users, and so on.
PHYSICAL SECURITY. This clause would address real-world security at the provider's facilities: access protocols, number and qualifications of security guards, authorization of visitors, nature of ID badges, and so on.
ENGINEERING SUPPORT. This clause would specify the nature of human support -- chat, phone, email -- and the hours of operation (such as "24 x 365," meaning every hour, every day). Rapidness of responses may also be specified.
MISCELLANEOUS. Depending on the provider, other clauses might refer to uptime of load-balancers, size and uptime of persistent storage, responsiveness of DNS servers, frequency of server reboots, procedures for disaster recovery, and so on.
A provider's Service Level Agreement outlines the services to be provided, the metrics of measurability for each service, and any penalties and/or remedies should the level of service be insufficient. If a network provider, for example, promises 99.999% uptime of their network, customers should reasonably expect no more than 5.26 minutes of "failure" per year.
In addition to specific, measurable services, the Service Level Agreement will typically include clauses concerning support options, dispute resolution, disaster recovery and so on. The following are some of the services, terms and clauses contained in typical SLAs.
UPTIME/AVAILABILITY. This is a measure of the time a service is properly functioning. The uptime promise may be 99.9%, 99.999% or even 100%. Availability refers to the reliability of a system -- uptime compared to downtime. Different services -- computer servers, cloud storage, network operation -- may have different uptime guarantees.
CREDITS. A credit is any payment or waiver to the customer as compensation for insufficient uptime or some other service failure. In order to obtain a credit, you may need to provide your web host with proof of the down time. There are a lot of different monitoring companies that you can use. Most of them require a paid subscription, but if you need to be monitoring your SLA, it might be worth it. CopperEgg (www.copperegg.com), Manage Engine (www.manageengine.com), and Dotcom Monitor (www.dotcom-monitor.com/sla-management.com/aspx) are just a few of the more well-known options, but there are a lot of other good companies too.
PERFORMANCE METRICS. Such clauses would numerically specify the characteristics of service sub-components such as internal/external network latency and throughput, hard drive I/O rates, storage quotas, CPU speed and availability, maximum simultaneous users, and so on.
PHYSICAL SECURITY. This clause would address real-world security at the provider's facilities: access protocols, number and qualifications of security guards, authorization of visitors, nature of ID badges, and so on.
ENGINEERING SUPPORT. This clause would specify the nature of human support -- chat, phone, email -- and the hours of operation (such as "24 x 365," meaning every hour, every day). Rapidness of responses may also be specified.
MISCELLANEOUS. Depending on the provider, other clauses might refer to uptime of load-balancers, size and uptime of persistent storage, responsiveness of DNS servers, frequency of server reboots, procedures for disaster recovery, and so on.