Val and 'Zilla were having one of their famous love-ins on Twitter with regards to support of EMC arrays behind the v-Series. It's a topic which has a habit of raising it's ugly head with fairly monotonous regularity.
The NetApp vSeries does not appear on the EMC Support Matrix (one of the few things which doesn't; last time I checked the Goblin Teasmaid was listed on page 976) and hence, it is true to say that it is not fully supported by EMC. And this is where at times it gets contentious and it all comes down to the definition of support.
If you ask EMC whether a DMX will work behind a vSeries, you will get the answer 'It's not supported' but that's not the question you asked. Even if you ask the question is there any reason why a DMX (or CX or VMAX) won't work behind a vSeries, you will get the same answer! If you keep pushing, often you will get dark mutterings that it should work in most cases but there may be border-cases where it won't work and EMC won't support you!
But NetApp can point to a number of customers where it does work and a number of customers where EMC might even do more than take the stance 'It's not supported, if it goes wrong, it's your problem!' I might even have a vague idea where some of these customers are (not us BTW).
The simple truth of the matter is that EMC will support their array and if no fault is found with the array; then that's where in general (unless you are a super special customer), their support will stop. NetApp will support their head and will do their damndest to make it work with EMC arrays. EMC, as far as I know are not doing an Apple and detecting that a NetApp head is talking to their array and then sulk a la Apple, Palm and iTunes!
And I suppose it comes down to access to engineering teams; EMC can work with the other operating system vendors quite happily and get access to the appropriate bits of source and vice-versa but I suspect the same isn't true when dealing with NetApp. Perhaps it could be made true, not sure? But as a customer, it's damned annoying! A v-Series in front of CX for example is potentially an excellent combination. EMC don't actually know whether their arrays will work behind a NetApp head because they will claim that they can't know.
BTW, I believe that the same is true for EMC arrays with SVC and USP-V. EMC will not officially support anyone else's storage device in front of their arrays (I am prepared to be corrected on this mind you!).
I guess the solution is for NetApp to buy EMC? Or the other way round? Or perhaps Cisco could buy them both and bang their heads together.
p.s FWIW, evidence suggests that EMC arrays work quite will behind NetApp heads! And, oh yes; the Celerra VM works pretty well with NetApp storage behind it.
Just this week our EMC maintenance renewal quote came and there was a big yellow note next to the CX700 line items:
“EMC IS DETERMINING IF THEY WILL SUPPORT THESE ITEMS WITH THE V SERIES ATTACHED”
It will be interesting to see how this plays out.
It does seem silly, like EMC discontinuing the Data Domain Gateway product too. Why mess with special one-off support deals, just replace the end of life EMC system with something that is supported by Netapp? As a 3PAR employee, I recommend a 3PAR T Class or F Class system and we will very happily support Netapp’s V series and vice versa. If a customer uses our new Thin Conversion technology they will probably also end up with much higher utilization too.
My previous comment about EMC discontinuing the Data Domain Gateway product was obviously wrong. EMC has NOT discontinued the Data Data Domain Gateway product (as far as I know). What I meant to say was that the future support matrix (and quite possibly the current support matrix) for the Data Domain Gateway could be smaller than it was a few months ago. Mea Culpa.
Any time this comes up and some one says they want EMC’s support for such a configuration when I ask why the answer is the same.
“EMC’s RAID back end is more reliable than NetApp’s”
Were it not the case people would forego the V-Series entirely and just buy a FAS.
A question for Marc Farley. What percentage of your joint customers are using the V-Series for NAS only and using the InServe for block?
That’s probably because they want to make use of the investment they made on the EMC storage instead of throwing it away and replacing it with FAS.
I was going to say the same thing. It may also be that the customer has gone through a long and tortuous evaluation process and have concluded that
1) EMC make the best block arrays
2) NetApp make a much better head
And together they are the best of breed solution. Performance with a DMX behind a NetApp filer is allegedly excellent and some customers may have deployed this configuration. NetApp are happy with the situation because they have a new customer, EMC are not happy with the situation even tho’ they are still the majority of the storage revenue in that customer.
So there you have it:
Netapp happy
EMC not happy
Customer not happy
Zilla, I don’t have the actuals in front of me (and don’t expect me to post them here), but a decent percentage of our customers are using our arrays with V series heads. They are defintely not one-off sales and the two companies have worked together to support customers.
Its not like Netapp is some Tom, Dick and Harry company. What would happen if EMC added V Series to the support matrix, would it really hurt EMC’s sales? Would Celerra sales dip? Probably. Enough for EMC to kill Celerra? Probably not. But would customers like it. Definitely.
I know its not the same situation as it is with VMware, but customers really like it when their products integrate well with each other. So what is happening with Data Domain? Is the gateway product going to remain open to the rest of the storage world or is it going to be closed off to EMC strategic interests? It’s not like Data Domain hasn’t already done all the testing.
Marc, last month I asked for a Data Domain gateway quote and was basically told no. The Data Domain SE said they are only approving new gateway configurations in special situations. I understood that to mean big accounts, which we are not (only eight DD’s in production).
Hi Martin –
I am the General Manager for the V-Series business unit at NetApp. Thanks for the comments. Our support of EMC products is always a hot topic especially for EMC customers who want the power of NetApp’s Data ONTAP storage operating system, but don’t want to change out their back end disk arrays.
First – let me say that we extensively test and certify V-Series with EMC arrays. We publish a support matrix that contains all supported configurations. So if an array is on the matrix, it has been tested. It is a rigorous, multi-month test process.
Technically, it is easy to understand how we can certify a third party array. V-series just looks like any other host to an array. The array simply presents LUNS to V-series. So most of the testing we do is to make sure we handle failures of an array controller, switch, or V-series head, among other items, correctly.
For support, if a customer has a support problem that impacts their hosts, they call NetApp. We can quickly determine if the problem is ONTAP related, interop related, or due to the back back end array. In the first two cases, we handle it. In the last case, the array vendor would work it. We have many happy joint EMC and NetApp customers. Just like NetApp is committed to customer satisfaction, that has been our experience with EMC and our joint customers.
We would enjoy a more formalized support process with EMC, but to date, they have been unwilling to support this. Despite this, customers should be feel comfortable deploying these solutions.
Bryan Semple
General Manager, VBU
Bryan, thanks for popping in! It’s good to hear that NetApp go through a rigorous qualification process and I’m sure that in most cases it is more than sufficient. I guess, the thing which worries most customers is if EMC release a firmware upgrade which breaks things from a NetApp point of view.
Fortunately, the process of certifiying the whole end-user data-centre environment against an EMC upgrade is so time consuming and then upgrading all the hosts to the required levels probably gives you time to certify and test your bits.
Hi Martin –
Firmware updates are generally certified within 90 days. So, yes, the datacentre process is more complex, so customers have plenty of time.
The more complex certifications are when a new platform ships with new controllers, ASICs, HBAs. These tend to be more time consuming first time through. But, market adoption for new arrays is sometimes slow. So this generally also provides a time buffer to deliver on a certification.
I can flip that Marc as I see after a few years of moaning from 3Par customers the company still hasn’t provided IBM with the information required to qualify 3Par gear behind SVC.
Barry Whyte saying that was exactly the problem a while back and lo and behold it’s still not on the SVC support matrix when I checked it tonight.
So there you have it:
3Par happy
IBM not happy
Customer not happy
As for the Data Domain gateway product I *cannot* speak for the Backup Recovery Systems division and will defer all questions about Data Domain offerings to BRS.
As we have found – we are happy to support EMC devices behind SVC. EMC not, because we show them for what they are… or whatever reasons…
When it comes to it, the customer is mutual, and if EMC turn round and say – its not supported – who are they going to beat up…
Last 24×7 call I was involved in I spent several hours explaining to the EMC support guy how a switch couldn’t possibly change the SCSI ID of a LUN and how there must have been an EMC port setting that was causing the problem – two days later we (IBM) solved what EMC seemed in-capable of working out and the customer has since moved from EMC to IBM storage.
My point is that in this day an age, we all have to be open, realised that customers will implement what they need and we all have to do a better job at supporting a heterogeneous environment. Thankfully with SVC we have a good head start.
That door swinging both ways Barry. It took IBM close to *three weeks* to ship a mutual customer of ours a replacement disk drive for a drive which had failed inside a node in his SVC cluster pair.
Incompetence can be found everywhere.
Zilla,
Just another example of EMC arrogance, Asked a legit question around support and you somehow find away to start bagging out 3 PAR. I suggest EMC start focusing on what the customers are asking for and not centering on other vendors issues.
If you want to throw stones, then you should check and see if your in a glasshouse or not.
Im sure that there are people out there that could comment on EMC’s lack of support for their customers. In fact, the 2nd biggest mining company in the world who i used to work for moved away from EMC for 2 reasons. EMC hardware was flaky as and had continual performance issues and the support we got from EMC was borderline disgraceful. Needless to say they now run NetApp in all regional sites across the globe and NetApp in front of their HDS storage which sits in the CSC Data Centre.
I do however realise that VCE is going to be the saviour of EMC, No new tool sets just another way to sell disks.
As one of my collegues once said. You Cant polish a turd
Paul where you always a coward or was it something you grew into?
I ask as you can’t even link to your own Twitter stream.
Since you add zero technical value to any debate you’ve taken part in, fleeing when your arguments are shown to be not only without merit but also without consequence your bloviating is just that.
The usual hot air from both orifices at the same time.
If Marc wants to ask about gateway support then I think that’s a two way street. He’d agree as he believes in robust debate and he’s not a coward nor are most other people here.
Though not only is your comment worthless but experience shows people of your stripe are also worthless employees. Which explains why you’re no longer working for the second biggest mining company in the world.
Go back to hiding, it’s all you are worth.
Hi Zilla,
Once again, you have stooped below the rules of blogging and have decided to make personal attacks against people in your fellow industry.I would suggest in future that you keep your comments in line with the topic and provide EMC perspective on why its not supported rather then point fingers at IBM and 3PAR. It only makes you both yourself and your company look childish in front of millions of readers. Hardly the ideal footying for a Company who is touting VCE as the biggest change in the IT world.
Now to answer your company, yes i left the 2nd biggest miner in the world, but alas the reason i left was to work for the Biggest Mining company in the world.
I will also make sure that a copy of your message is circulated throughout the worldwide BHP IT community so that this attack on me is dealt so that BHP’s fantastic culture is not tarnished by the dealings of some EMC representives like yourself
Im sure the management of EMC who im sure read these blogs will interested to see your actions and more then likely discuss them with you privately.
I think the best thing for you to do now Zilla, is issue an apology to all. We all make mistakes and im only happy to put my hand up when i have.
cheers
Ummm, Paul…why are you linking to Zilla’s twitter feed? From my point of view; that looks like an attempt hide your identity? If you have a website/linked-in profile/twitter profile; it would be much better and a lot more open if you linked to that.
I’m happy for robust debate to happen on the comments on my blog but I would prefer it if people were open who they were. I mean there is a Paul S who appears to work for a NetApp reseller in Australia in the area where your IP address appears to come from, who according to their Linked Profile once worked for Rio Tinto…hopefully this is a coincidence.
For all his faults ‘Zilla is always open as to his affiliations…wish everyone was!
i dont have a twitter account although i have now set one up. Vstoragemonger.
feel free to add me
Paul S. I just looked for Vstoragemonger on Twitter and couldn’t find you.
Zilla, I suppose we probably have had customers who wanted us to certify with SVC, but so far I’ve only heard it from you. Barry Whyte, is there an axe to grind here?