my company is in the process of migrating to SAN. we are working with Dell to provide hardware but i would definitely prefer using Cisco switches. can anyone give me a general idea of what are the requirements for switch to be "iscsi optimized"? also any recommendations on Cisco switch that would work for the entry-level SAN solution?
There are couple of things to consider with iSCSI.
- The NIC in the host...will you use a dedicated NIC for iSCSI or the same NIC that is doing other LAN connectivity so that iSCSI has to compete for NIC resources. There are some NIC optimized for iSCSI that are dedicated and can not be used for other LAN traffic.
- As for Gigabit Ethernet switches, I would recommend one that is guaranteed to be wire rate between the ports. The last thing you want is an oversubscribed switch impacting your storage access.
Look at the network design and watch out for over subscription. IE: If you have 5 isCSI hosts attached to switch 1 all at 1GB. While the iSCSI targets are on Switch 2, and there is a 2 GB trunk between the switch 1 and 2, it might pose a problem.
Jumbos would be a good option, but the NIC and iSCSI target would need to support Jumbo for you to realize the benefit. Other than the over subscription and jumbo support, there is nothing unique about an iSCSI optimized switch that I can think of.
If you are serious about iSCSI you should research the network design, and redundancy issues that come into play. I know that Linux has built in multipathing that works quite nicely. I believe that the latest Microsoft iSCSI drivers also support muiltipathing. This would require 2 NICs for iSCSI and a redundant network design.
If you are using an iSCSI gateway to access fibre channel devices for storage, look into how you can provide redundancy and fail over on the Fibre Channel side as well.
Your initial post is over a year old, so you have probably completed your SAN integration.
But let me offer you this.
The Dell 6248 offers iSCSI optimization in the form of QoS prioritization for iSCSI packets that share the same NIC. With a wire rate, non-blocking switch with 1:1 oversubscription and 10G connectivity, contention for data processing and forwarding resources will probably be less of an issue. iSCSI has no problem competing for resources in a converged fabric when even a 2:1 or 3:1 oversubscription rate exists, especially with 10G deployed.
Cisco's answer to iSCSI is FCoE which wont only be enhanced by 10G ethernet, but will actually need it to function properly. It will also need the CEE -- or what Cisco calls DCE -- enhancements to ethernet which will create a unified fabric that is lossless and can simulate a dedicated FC fabric.
VMware Trunk Port Group is supported from ACI version 2.1
VMM integration must be configured properly
ASA device package must be uploaded to APIC
ASAv version must be compatible with ACI and device package version
In the Previous articles of ACI Automation, we are using Postman/Newman as the Rest API tool to automate the ACI Configuration.
In this article I’m going to discuss on usin...
One of the first steps in building your ACI Fabric is to go through Fabric Discovery. While Fabric Discovery is usually a straightforward process, there are various issues that may prevent you from discovering an ACI switch. This article wil...