No Single Point of Failure Configurations

All Luminex Channel Gateway solutions can be configured with No Single Point of Failure.  The fundamental requirement is to insure the ongoing operation of the solution in the event of a subsystem failure.

There are two primary components to each Luminex Channel Gateway solution, the Channel Gateway server and the storage device.  There may also be supporting components within the server, the storage device and the FC/Network topology that provide redundancy, multiple paths, clustering and automated failover.  Recovery timeframe requirements and cost constraints control the extent to which redundancy and recovery automation are proliferated throughout the environment.

Standards-based and Solaris-based Channel Gateways

The Channel Gateway is a Solaris-based server installed with the Luminex Channel Gateway software and mainframe channel interface cards, either ESCON or FICON.  In addition to the Luminex software, Sun software products are installed to provide the desired level of fault tolerance, failover performance and high availability.  In all configurations, the Channel Gateway servers do not create a single point of failure.

Uninterrupted Access to the Storage Device
 
The Channel Gateways deliver a no single point of failure access to the storage device. Redundancy exists throughout the solution and all data is available simultaneously from all of the Channel Gateways. 

For NAS Storage environments, the NFS protocol, along with Luminex’s file locking mechanisms, combine to deliver uninterrupted access to the storage device.

For Fibre Channel, SCSI and iSCSI Storage environments, the Sun Cluster, Sun QFS, Sun Volime Manager and Sun mpxio software products are deployed alongside the Channel Gateway to deliver uninterrupted access to the storage devices.

Storage Devices

A wide variety of storage devices are supported by the Channel Gateway.  In fact, nearly any NAS/NFS, Fiber Channel, SCSI or iSCSI storage device is supported.  But to achieve a no single point of failure configuration, the use of a sufficiently configured high availability or fully redundant class of storage system is required.

Recovery

Recovery after a failure may differ from one configuration to another – primarily due to the characteristics of the storage device.  Normally in the event of a component failure, the entire system continues to operate at a diminished performance level. Notification of the failed component is presented and maintenance can be scheduled for the device. The crucial feature of a no single point of failure configuration is that regardless of the failure, no data is lost, and the system can be operational at a diminished level until the failed component is repaired or replaced.

Most of Our Solutions Require No Database

The majority of our Channel Gateway configurations do not require a database.  The mainframe tape data is stored in a self-describing manner within a standard filesystem. This includes all Data Domain-based, Sun Storage-based, all NAS-based and all FC-based storage solutions. 

In addition, the Channel Gateway can be configured to support both physical and virtual tape libraries.  For this class of storage, such as VTL-based solutions from Sepaton and Copan or real tape libraries for Sun/Storagetek, a database is used within a service-oriented-architecture environment.  The Luminex solution uses an active-active and clustered database implementation to manage tape volume information.  Our active-active and clustered solution delivers true no single point of failure.  All data is always available. 

Suitability for Mainframe Applications

Luminex’s no single point of failure configurations are well suited to support mainframe tape operations.  The mainframe expects to use a specific collection of devices for certain jobs and should one of those devices be unavailable due to a failure, the mainframe will select the next available device from the collection.  Depending on the tape application being used, the job may be resubmitted to the available devices automatically or the operator might have to indicate that the failed devices are offline.

BACK TO TOP

Examples

Channel Gateway Servers with NFS Attached Storage

• Only the Channel Gateway software is required.
• All Channel Gateways can be used simultaneously.
• Shared access is provided by the architecture of the Channel Gateway software.
• If any Channel Gateway fails, the others will continue to operate and access all data without manual intervention.
• The affected mainframe channels will be boxed or varied offline.

BACK TO TOP

Channel Gateway Servers with FC Attached Storage

• All Channel Gateways can be used simultaneously via Sun’s QFS software that provides shared multi-port channel access.
• Solaris Volume Manager mirrors writes to each FC array
• If the primary QFS metadata server fails, the backup metadata server is automatically promoted to the master metadata server role by Sun Cluster. 
• The affected mainframe channels will be boxed or varied offline.

BACK TO TOP