As Fibre Channel Storage Area Networks (SANs) continue to serve as the backbone for reliable delivery of business-critical applications, SAN storage volume replication has become the prevalent method for remote data replication. While SAN replication assures that information assets are protected in the event of a disaster at the primary site, remote replication can be quite complex, involving short distance synchronous replication in combination with very long distance asynchronous replication across one or more WANs. Additionally, administrators of today's SAN replication infrastructures face "fabric blindness", the inability to accurately plan and implement remote SAN replication as current SAN monitoring processes do not provide the deep visibility into data paths or the protocol-level analysis capabilities on a per volume basis required to track and monitor high-availability replications. This presentation will provide details on various replication scenarios, applicable metrics and faults to monitor, and actions that administrators can proactively take to mitigate Fibre Channel replication failures. Key topics include:

• The differences between SAN replication and IP-based replication (for example, there are normally no intermediary servers between replicating storage volumes in SAN replications)
• How to configure non-intrusive, embedded probing of storage traffic to detect faults and create meaningful metrics for analysis
• Determining absolute values of latency, read and write completion times, throughput speed, and percentage utilization of links through correlation of every packet of each replication I/O exchange
• Using alarms to trigger collection of more information or notify administrators when performance metrics exceed acceptable thresholds
• Filtering of non-replication related traffic to determine its impact on the replication process
• Developing proactive processes surrounding critical replication related scenarios, including detecting replication slowdowns at the volume level, diagnosing SCSI protocol faults related to replicating volumes, detecting long distance bottlenecks and identifying whether it is storage subsystem or WAN related, identifying device related failures, and analyzing any non-replication or low level fibre channel protocol activity that could or will eventually impact replication performance

Dave Simpson Editor-in-Chief InfoStor
Harish Nayak Vice-President Enterprise Marketing

Finisar Corporation (NASDAQ: FNSR) is a global technology leader for fiber optic components and subsystems and network test and monitoring systems. These products enable high-speed voice, video and data communications for networking, storage and wireless applications over Local Area Networks (LANs), Storage Area Networks (SANs), and Metropolitan Area Networks (MANs) using Ethernet, Fibre Channel, IP, SAS, SATA and SONET/SDH protocols. The Company is headquartered in Sunnyvale, California, USA. More information can be found at:  http://www.finisar.com