ADVANCED MANAGED SERVICES PROVIDER

High Availability Auto-Failover and Disaster Prevention

 

Disasters come in all types and degrees of severity

There are many variations of Disaster Recovery services which can be deployed to meet your specific requirements.

High Availability and Auto-Failover

• Automatic Failover
• In Bound Load Balancing
• Eliminates DNS Propagation delay
• Multi-Data Center Deployment 
• Typical Failover within 90 seconds

CPAC deploys a hosted service that uses dynamic DNS to achieve fault tolerance for Internet bound servers. The commercially proven Domain Name Server infrastructure is a redundant, fully meshed world-wide infrastructure.
 

Auto-Failover

This is the next generation for Auto-Failover and load-balancing which raises the bar for fault tolerance and high availability for all Internet bound severs.

Disaster Recovery - Disaster Prevention

Deployed as a service, HA provides unprecedented uptime for your servers and your business. HA can be used to automatically failover from primary to secondary servers. For example, in the case of 2 servers or server farms, a typical implementation may have one server defined as primary or active and the second as backup.

When the primary server becomes unavailable, HA identifies the outage and immediately changes the appropriate DNS records to reflect the correct IP address of the backup server/s. This change is then propagated across the entire NET DNS infrastructure in seconds and then to the entire Internet. Total propagation time can be as low as 30 to 45 seconds.

From single server environments to multiple geographically dispersed cluster farms, HA provides fault tolerance on the Public side of the Internet.

CPAC’s partner network maintains six redundant DNS server sites located across North America and Europe. These sites are connected through six different major Internet backbones. This ensures that all DNS records are protected against any geographic disruption and against the failure of any one primary Internet backbone.
 

HA provides for various notification mechanisms

For any HA action, many notification options are available. HA can also be used to load balance across the 2 links.

HA makes use of MON, a newly designed monitoring suite made up of the following features:

MPM, Multi Point Monitoring. MPM allows for the definition of 1, 2, 3 or more monitoring locations all testing the availability of a defined server. Multiple monitoring points (MP’s) are crucial for highly available environments.

Only with MPM can false positive server down messages be correctly identified. With MPM server down messages need to be reported from all participating monitoring locations, in order for the Auto-failover action to initiate.

PMA, Persistent Monitoring Agent, PMA actually enables the regeneration of any individual monitoring point. With these new capabilities, if any type of monitor agent failure occurs or if any problem at a particular monitoring location is encountered, PMA will automatically regenerate each of individually effected MON monitoring agents at a pre-defined alternate NET location.
 

Domain Name Server Infrastructure

The DNS system is proprietary and is not susceptible to the same attacks that have disrupted DNS servers running other DNS software.

Scalability

HA can support up to 10 different server locations. The locations can be geographically dispersed and can be connected through different backbones. HA can also support a wide array of server configurations at each of these locations. There are also many options that can increase HA's functionality.

Load Balancing

HA can also be configured to provide load balancing between a number of servers. Even if they are in different geographic areas, HA can distribute the traffic across the server farms.

Weighted Server Priority

You can specify that each server handle a certain percentage of server requests, directing more requests to the more robust servers.

"Intelligent" round robin

When a server fails, HA removes it from the rotation however it continues to monitor all the servers. When the server is back online HA automatically returns it to the rotation and restores your originally specified weight to each server.

Servers may be in geographically dispersed areas

By placing the servers in different geographic areas you can protect against regional disruptions yet continue to balance the load among the active server locations.

Using different Carrier backbones for your servers reduces the risk of failure due to the failure of one or multiple Internet backbones.
 

HA complements existing load balancing devices

If you're using a load balancing device at one or more of your installations, HA can balance the load among your locations, while your load balancing device can balance the load within each location. Using HA to augment your hardware load balancing devices protects against failure of the load balancing device and protects against the loss of connection to one or more of your installations.

Customizable Actions

Action is taken ONLY if both/All monitoring agents report a failure.

At each polling interval the agents compare results and takes action only if both/all agents detect a failure. Redundant monitoring prevents a notification and a failover from occurring unnecessarily. 

Protocol Monitoring

HA can monitor different protocols, HTTP, HTTPS, SMTP, FTP, TELNET, PING (ICMP) and POP3. By request other protocols can be added. If HA detects a failure by monitoring one of these protocols, you have the options of failing over just that protocol or failing over the entire domain.