SHARED VS LOCAL STORAGE

Download The decision for system architectures base on the needs of the system now and the foreseeable future. Single points of failure are unaccept...

1 downloads 897 Views 4MB Size
SHARED VS LOCAL STORAGE The decision for system architectures base on the needs of the system now and the foreseeable future. Single points of failure are unacceptable in a world class manufacturing environment, so high availability features need to be built in from the beginning. Sometimes the required components to provide shared storage are very different than what might traditionally be seen in legacy automation environments and might seem a bit scary or be misunderstood as overdesign. It is essential for any system requiring high availability to have shared storage. The only time a local storage only solution is viable is if the process and system do not need to scale and grow and if the servers running can tolerate a failure. Often, local storage only solutions are found in smaller OEM skid type standalone applications. Once you get into the connected enterprise world, distributed architectures, high availability, and Industrial Internet of Things (IIOT), having shared storage is essential to providing an infrastructure level solution to support the needs.

SHARED STORAGE

LOCAL STORAGE

ADVANTAGES

ADVANTAGES Faster performance

Consolidated Platform (centralized) Redundant storage controllers

Cheaper direct costs

Scalable/expandable Can add more hosts on the fly and access same storage / VMs System can tolerate failure of a host

Work well in systems that only need software failover Simpler/less configuration - less experience/expertise needed

Allows for High Availability within VMware - Ability to move VMs between hosts - Ability to perform maintenance on a host without bringing system down

DISADVANTAGES Higher costs More experience/expertise needed

DISADVANTAGES Can't move VMs between hosts System can't tolerate failure of a host Rely on local non redundant RAID controllers

More cabling Not as fast performance as direct attached / local

Not as scalable or expandable Does not work well with systems with SQL

-since that can't tolerate software failover

For more information visit www.cybertrol.com or call us at (763) 559-8660