Thank you for taking the time to share information with the community. there’s nothing like an open dialog.

I believe you have one error in your post. When you claim you can’t compare a unified array to a SAN. This is untrue, just compare the NetApp SAN to a Symmetrix or a Clariion. This allows cusomters to understand the differences in storage savings technologies.

I went into details here:

http://blogs.netapp.com/virtualstorageguy/2010/07/myth-busting-storage-guarantees—part-ii.html

Again, thanks for sharing. If you have any feedback, just ping me. I’m more than happy to advance my knowledge base.

1) Sub LUN auto-tiering (FAST)

2) Ability to failover without losing access or requiring re-warm to extended cache (EFD in our case and PAM in ntap’s).

3) Ability to add ports on the fly with no outage (Ultraflex)

4) Integrated I/O splitting for CDP

5) N+1 Clustering for NAS without breaking other mainstream features like snapshots and replication

6) Dynamic LUN migration – (move a lun to any place RAID type or tier in the array without an outage)

7) Ability to manage the array through a single console (not a single console opened for every controller)
Space guarantee offering that doesn’t require 96 pages of caveats with things like reduction of databases and other common files to no more than 10% of the total data set.

9) Ability to extend cache with EFD for both reads and writes (and no, nvram isn’t always the answer when back to back CPs happen)

I could keep going but this is probable enough for now

Further, you can’t edit a config file directly either, you will need to use “rdfile” to display the current file, then edit it in another tool on your desktop, and use “wrfile” to write a new version of the file and paste the edited copy from your desktop.

There are other, more advanced methods, you can use though. NetApp has a PowerShell tool for Windows savvy admins, or you can use Perl to script things and possibly leverage RSH to run remote commands against the filer. Hope that helps.

All in all your re-representation of the chart is valid, but really just re-arranges things and cleans things up because you took issue with the way another person represented the information. You’re both correct in your own views of the data and the way you rearranged the chart isn’t necessarily demonstrating anything ground breaking or proving something false which another claimed to be true.

I think your headline is a bit sensationalist since you’re not actually calling anybody out on any lies that they’ve put forward. Instead you’ve just performed some clean-up and rearrangement of the data.

I only take issue with two of your footnotes and they’re pretty minor nits. I’m assuming your 2nd footnote is for the replication category, but your chart is missing the label. Also, you begin with RecoverPoint CRR and end stating just RecoverPoint. NetApp’s SnapMirror does compete/have the same features as a lot of the RecoverPoint pieces, but certain aspects of CRR or Application-level CDP don’t exist in their bag of goods. Both products perform asynchronous, semi-synchronous, and fully synchronous (restrictions apply) replication with data compression (again, restrictions apply), but there are certain RecoverPoint pieces that do go above and beyond in their functionality than what NetApp has to offer with SnapMirror and their SnapManager products as far as I know. Personally I’m a fan of both products for different reasons/applications.

In regards to the V-Series front-ending Symmetrix I think there’s a certain sanity check and common sense that applies. NetApp has their own HCL just like any other vendor and if you upgrade to something that isn’t currently on that HCL and supported then you obviously run the risk of running into issues. If NetApp says that version of code and that hardware is qualified, then they’re your point of contact for support – not EMC – assuming you have a support contract with them. In my opinion (so take it as you will) EMC never has to support a V-Series or any other third-party gateway device in front of any of their boxes if the other organization is going to offer you full support on it.

On a side note, at the end you talk about how front-ending a freight train with a box truck might not be in the end-user’s best interest. One would hope the end-user and the vendor, NetApp or other, have performance metrics to base a configuration on so those performance expectations are met before putting something in production. One could also argue that with the advent of virtualization and databases supporting NAS protocols along with larger Ethernet pipes that the market’s need for freight train SAN storage is becoming less and less important. That’s probably more for another discussion though…

Gregg & Dannae