Offloaded Data Transfer (ODX)

[pit]

Do you plan to implement this feature in near future?

regards,
Lukas

[anton (staff)]

Keeping in mind MS had basically failed to chew away noticable market share for hypervisors - not very soon. Fully functional and certified VAAI would be first (will represent next iteration of it with V8).

Do you plan to implement this feature in near future?

regards,
Lukas

[PCPSTechnology]

We are going to go to Server 2012 hyper-v this summer and the Hyper-V ODX would be of a great benefit to our institution. I would think that you would want to support this because while hyper-v may not have dominate market share, they are starting to make major inroads on the value side of virtualization (even 2012 Hyper-V free server has live migration and fail-over which ESXi does not.) Having a value added solution to work with Hyper-V is the main reason we went with Starwind.

As your VAAI is coming out next month, will you then start work on Hyper-V ODX? If not, will Starwind iSCSI have any significant benefit over a Windows 2012 file server with SMB 3.0 and storage spaces, as you can now multi-home and multi-path to a file share with enough resilience that both cluster shared volumes and SQL failover clusters are fully supported?

[anton (staff)]

1) We'll proceed with ODX after we'll release V8 with VAAI support. ESXi is dominating the market so it's a higher priority for us (ignoring the fact we're native Windows app
and not VMware kernel module).

2) StarWind is still ages faster (reads go local to SAS instead of touching network with SMB 3.0, writes are confirmed once over network and SMB 3.0 needs double I/Os here, add in-line dedupe for StarWind reducing amount of data actually touching the spindles Vs. increased amount of data in case of MS off-line dedupe, add flash and RAM caches for StarWind as MS has no flash cache and can only cache reads but not writes with CSV cache) and provides smaller installation footprint (only two or three Hyper-V running servers instead of a at least two Hyper-V servers + pair of Windows Server 2012 boxes to present SMB 3.0 share + SAS JBOD to feed shared storage to Clustered Storage Spaces).

We are going to go to Server 2012 hyper-v this summer and the Hyper-V ODX would be of a great benefit to our institution. I would think that you would want to support this because while hyper-v may not have dominate market share, they are starting to make major inroads on the value side of virtualization (even 2012 Hyper-V free server has live migration and fail-over which ESXi does not.) Having a value added solution to work with Hyper-V is the main reason we went with Starwind.

As your VAAI is coming out next month, will you then start work on Hyper-V ODX? If not, will Starwind iSCSI have any significant benefit over a Windows 2012 file server with SMB 3.0 and storage spaces, as you can now multi-home and multi-path to a file share with enough resilience that both cluster shared volumes and SQL failover clusters are fully supported?

[PCPSTechnology]

1. Glad to hear that it will be in the works.

2. I figured as much, but wanted to be sure. I was going to put both our StarWind iSCSI Enterprise HA side by side with a SMB 3.0 file share and the same Windows 2012 box and do some stress tests, but I will look for that article to see if they have the metrics I need (always have to justify expenditures with cost/benefit analysis.)

Thanks for the info.

[Anatoly (staff)]

Great to hear that your questions was answered here.

I`ll ask you to keep us posted.

[anton (staff)]

You should proceed with vSAN (ex-Native SAN for Hyper-V) and not core product running on bare metal.

1. Glad to hear that it will be in the works.

2. I figured as much, but wanted to be sure. I was going to put both our StarWind iSCSI Enterprise HA side by side with a SMB 3.0 file share and the same Windows 2012 box and do some stress tests, but I will look for that article to see if they have the metrics I need (always have to justify expenditures with cost/benefit analysis.)

Thanks for the info.

[anton (staff)]

V8 Beta with VAAI is out. We'll do our best to have ODX supported with V8 release.

[awedio]

2) StarWind is still ages faster (reads go local to SAS instead of touching network with SMB 3.0, writes are confirmed once over network and SMB 3.0 needs double I/Os here, add in-line dedupe for StarWind reducing amount of data actually touching the spindles Vs. increased amount of data in case of MS off-line dedupe, add flash and RAM caches for StarWind as MS has no flash cache and can only cache reads but not writes with CSV cache) and provides smaller installation footprint (only two or three Hyper-V running servers instead of a at least two Hyper-V servers + pair of Windows Server 2012 boxes to present SMB 3.0 share + SAS JBOD to feed shared storage to Clustered Storage Spaces).

Are you saying Starwind ISCSI is faster (therefore preferred) for Hyper-V compared to multi-channel SMB3.0?

[anton (staff)]

Yes, absolutely. Even with R2 Microsoft still re-directing I/O to the "preferred" SoFS which is current LUN owner. With StarWind everything is completely independent.

http://technet.microsoft.com/en-us/libr ... 31474.aspx

--

Feature/functionality New or updated Summary

Automatic rebalancing of Scale-Out File Server clients

New

This functionality improves scalability and manageability for Scale-Out File Servers. SMB client connections are tracked per file share (instead of per server), and clients are then redirected to the cluster node with the best access to the volume used by the file share. This improves efficiency by reducing redirection traffic between file server nodes. Clients are redirected following an initial connection and when cluster storage is reconfigured.

--

2) StarWind is still ages faster (reads go local to SAS instead of touching network with SMB 3.0, writes are confirmed once over network and SMB 3.0 needs double I/Os here, add in-line dedupe for StarWind reducing amount of data actually touching the spindles Vs. increased amount of data in case of MS off-line dedupe, add flash and RAM caches for StarWind as MS has no flash cache and can only cache reads but not writes with CSV cache) and provides smaller installation footprint (only two or three Hyper-V running servers instead of a at least two Hyper-V servers + pair of Windows Server 2012 boxes to present SMB 3.0 share + SAS JBOD to feed shared storage to Clustered Storage Spaces).

Are you saying Starwind ISCSI is faster (therefore preferred) for Hyper-V compared to multi-channel SMB3.0?