Don’t get in a cluster fluster, Wikibon tells NetApp users The Wikibon consultancy, in what amounts to a sustained analytical assault on NetApp’s product strategy, claims NetApp 7-Mode array users shouldn’t update to CDOT (Clustered DataONTAP), its latest FAS array operating system, and should think seriously about moving workloads to other suppliers’ systems. CDOT clusters FAS arrays …
One minor nit:
Wikibon also says true private clouds need to have chargeback which is, candidly, bullshit. Chargebacks are an accounting function, not a cloud operations function. Chargebacks are also not recognized inside of NIST's definition of cloud either: http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf.
What if an org doesn't need chargeback? Is it not a cloud then? It's a silly argument.
Again, a minor nit but it is an example of how "one true cloud" definitions don't always work.
This article definitely can't be biased towards Oracle. "For the cheapest and best private cloud option, run Oracle engineered systems" "For the best NAS options run Oracle ZFS" "back up to the Oracle cloud!" In what freaking world is an Oracle engineered system going to be the cheapest?
Listen I resell NetApp and still think they have great components to great solutions (flexpod) and my customers would agree so maybe I'm biased but this article is so pathetic it's entertaining.
Hi All, Dimitris from NetApp here.
It is important to note that Mr. Floyer’s entire analysis is based on certain very flawed assumptions. Here is some more detail on just a couple of the assumptions:
1. Time/cost of migrations:
a. The migration effort is far smaller than is stated in the article. NetApp has a tool (7MTT) that dramatically helps with automation, migration speed and complexity.
b. It is important to note that moving from 7-mode to a competitor would not have the luxury of using the 7MTT tool and would, indeed, result in an expensive, laborious move (to a less functional and/or stable product).
c. With ONTAP 8.3.2, we are bringing to the market Copy Free Transition (CFT). Which does what the name suggests: It converts disk pools from 7-mode to cDOT without any data movement. This dramatically cuts the cost and time of conversions even more (we are talking about only a few hours to convert a massively large system).
d. NetApp competitors typically expect a complete forklift migration every 3-4 years, which would increase the TCO! Mr. Floyer should factor an extra refresh cycle in his calculations…
2. Low Latency Performance:
a. AFF (All-Flash FAS) with ONTAP 8.3.0 and up is massively faster than 7-mode or even cDOT with older ONTAP releases. To the order of up to 3-4x lower latency. cDOT running 8.3.0+ has been extensively optimized for flash.
b. As a result, sub-ms response times can be achieved with AFF. Yet Mr. Floyer’s article states ONTAP is not a proper vehicle for low latency applications and instead recommends competing platforms that in real life don’t perform consistently at sub-ms response times (in fact we beat those competitors in bakeoffs regularly).
c. AFF has an audited, published SPC-1 result using 8.3.0 code, showing extremely impressive, consistent, low latency performance for a tough workload that’s over 60% writes! See here for a comparative analysis: http://bit.ly/1EhAivY (and with 8.3.2, performance is significantly better than 8.3.0).
So what happens to Mr. Floyer's analysis once the cost and performance arguments are defeated?
Thx
D
Who can match NetApp's file services capabilities? I moved off a celerra a few years back over to NetApp and I've got 19 vFilers that aren't going anywhere. Is there any file product on the market that can match NetApp's CIFS/NFS capabilities as well as vfiler dr feature?
I'm just curious how much Oracle paid him to shill that? Seriously, the Oracle ZFS appliance as a cheaper replacement? There's a reason they can't give those things away, and it's not because they're cheap, effective, or in any way "good".
I hope he's getting paid per click and doesn't actually care at all about his credibility, because this is pants on fire insane.
It's only if you use the constraints that are in the paper that Oracle has a prayer. Exactly those conditions, no less. {Shrug} It was sort of interesting examining the background articles but I'm a database geek at heart. Then again, I was using Intelligence Engineering as a category in my ontology here long, long ago which is where the tools that they are citing would be useful. Which is where/why they are talking about Systems of Record and Systems of Intelligence caught my attention. If you aren't a large (20,000+ employees) enterprise, this paper isn't relevant to what you do. If your "big data" needs are already covered, this isn't for you. Etc.
We have migrated almost all our 7-mode filers to C-DOT. (approx. 42 heads).
The last I checked, we had 100% uptime across all clustered systems and we now never need to negotiate a storage outage with business units for maintenance. Our life prior to C-DOT was definitely not 100% up and had an awful low of negotiation for outages. We've also designed the systems well so don't see performance issues on any of our critical system....
Not sure about you, but I think that's a pretty compelling reason to upgrade.
Say it ain't so Chris - the TCO argument in this article is totally contrived. So many errors, so little time:
1) Cost of 16 month cutover from 7-mode to C-mode system ($622,000). Really? 16 months for one storage system? In my experience 1 month would be more than enough time.
2) Cost of planning, upgrading skills, and testing ($275,000). Unique to C-mode? Nope - upgrading to any new storage platform, even ZFS, would incur this cost.
3) Additional cost benefit of C-Mode Migration: How about nondisruptive operation? Zero planned downtime? Seamless scale out and scale up? Access to a universal data fabric? Multi-cloud data mobility? Hmm, there must be some $$ value there too.
IMO you can take at least $1M from the Wikibon 6-year C-Mode costs and show a payback of 12-24 months. Anyone that wants to do a more realistic TCO analysis - ask your Rep or SE to run a TCO report @ www.netapprealize.com.
Cheers,
Larry@NetApp
If I could take a shotgun to my old 7 mode filers I would. Once you get cdot up and running (and if your migrating from 7 mode, the 7MTT makes it relatively pain free with cut over times in minutes per application.
There is not a lot of cost advantage to replacing a disk system early/ regardless of whose kit you run and yes if your doing a technology uplift your going to have to buy a new system and migrate to it but your replacing your disk systems every 3 - 5 years anyway. The cost of migration is really one admin's salary for a few weeks and once your in cdot your workloads migrate as you replace filer heads.
The real fun stuff begins when you take add some AFF nodes into your cdot cluster and seamlessly migrate your users to flash without taking any downtime.
OK - let's feed the troll...
NetApp's Data Fabric helps real Cloud users on AWS & soon Azure:
- Accelerate their development workflows via faster thin clones for rapid CD and broader CI testing
- Simply and efficiently replicate across regions
- Minimize their storage costs via dedupe, compression, thin provisioning and rapid snapshots
- Minimize their network costs via dedupe & compressed replication that doesn't rehydrate
- Stay compliant with international data domesticity requirements
- Arbitrage between supported Cloud providers (for NPS also including GCE, SoftLayer & Oracle Cloud)
- Add enormous strategic business value to almost every industry vertical.
For a hands-on demo, see this excellent post: https://www.linkedin.com/pulse/building-data-fabric-aws-cloud-ontap-tony-lee?trk=hp-feed-article-title-share
When I first heard it, the Data Fabric story was really good, it was eye opening, this is the bridge between private and public could as well as multi-cloud.
As implemented so far it's half vaporware and half a series of hacks at a pretty hefty price point.
From my testing a lot of what Val/NetApp mentions is true, we could just never get the price to justify it and ran into strange roadblocks along the way. We tried cloud ontap on AWS, wow, we could see it get pricey. The plus side is the ONTAP overhead is only the file system and ONTAP itself so it was only about 15%.
So we looked at NPS, and yea super pricey, with the added bonus of getting more expensive. We were told he uncapped interconnect fees were about to go up in price and have bandwidth limits added.
If you're going to arbitrage between two public clouds, you're going to have to find some mighty hefty discounts or be huge to cover the two active interconnects, and if you're huge enough to arbitrage I don't think you're looking at spot rates, I'm pretty sure you've got some special pricing already.
The Data Fabric value, at least today, isn't deduplication and compression to save storage space or bandwidth, or snapshots, it's just being able to move data, which is, don't get me wrong, really valuable, The other things don't really matter because the bandwidth savings and storage savings never really make up for the cost, although thin provisioning is kind of useful to reduce pay per month.
I really like NetApp, I've been using it for well over a decade, I was an early adopter of Cluster mode and really like it, even though I think I cut myself on just about every sharp edge possible. NetApp just needs to get the Data Fabric practice, up to the Data Fabric vision, stat. Every vendor pretty much has a <vendor> Private Storage at this point and there's other software that can do the data replication.
Late to the party, but one big thing stood out to me: 17% IRR is 'anemic'?
Let's use the current US Treasury 5 year bond yield of 1.44% as the risk free rate of return, which is a pretty standard thing to do. The 10 year bond is 2.01%, but 5 years is roughly the same investment horizon as the 50 months quoted. 17% IRR is nearly 12 *times* 1.44%, which is hardly anemic.
Hurdle rates are arbitrary, and if they haven't been reduced since interest rates plummeted in the Time of Shedding and Cold Rocks, then whoever is setting them isn't doing their job properly.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
