Myself and Tony Asaro have had a bit of snit over the uniqueness of the USP-V; he opines that it is unique and I am right that it is not unique. In many ways, this comes down to Tony's opinion that the USP-V is unique because it is the only external storage virtualisation array which is Enterprise Storage. In his opinion neither the v-Series or the SVC are Enterprise Storage and hence do not compete with the USP, DMX and DS8K range. Also in SVC's case because it does not have it's own disk and simply virtualises external arrays; it is not a storage device (I'll leave that comment alone).
So what this really boils down to is what is Enterprise Storage? A couple of years ago, I probably could have sat down and told you what is and what isn't Enterprise Storage but now? I'm not so sure, I can list you some characteristics of Enterprise Storage but the problem is that pretty much all of the arrays from most vendors have those characteristics!
-
Highly Available – 99.99%+ available
-
Highly Scalable – Supports 500+ disks and supports many hosts attached
-
Highly Performant – Whatever that means
-
Non-disruptive upgrades – Internal code and hardware can be replaced/upgraded with no service outage
-
Supports multiple RAID Levels
-
Supports multiple disk-types and sizes within the array
Problem is, as I say pretty much every array from most vendors have these characteristics. So what actually is Enterprise Storage or is it entirely defined by the price you pay? Are some things simply too cheap to be classed as Enterprise Storage?
You see, I'm no longer sure and does it really matter? I suspect it matters alot to the Hitachis and the EMCs of this world but to anyone else? For the rest of us, it probably comes down to the eye of the beholder. Thoughts anyone?
Its a good question – and certainly from a scalability, reliability, availability (and multi external disk) support statement SVC meets or betters all of them.
I think most of the team here would be deeply offended to hear that SVC isn’t a storage product !!! lol
Look at your invoice, does the £, $, or € at the bottom start with a # followed by at least 6 zeroes? If so you have “Enterprise Storage”. 😉
You’re right though, the distinction between what qualifies as a “mid-sized” vs. an “enterprise” array has been blurred significantly. Like you said, aside from HDS and EMC does anyone really care?
I totally agree with you and this is a good place we’re in. You no longer have to fret about which data you can afford to lose access to or plain just lose. With modern storage subsystems, making the decision on which class of storage I need to buy comes down to feature, function, support, preservation of investment and, of course, price.
Though all these criteria can be cast aside if you golf with the right person.
Martin
I’d add on something relating to the impact of component failure. Enterprise arrays tend to suffer less impact on component failure (e.g. cache board, an internal controller etc) compared to modular devices. This isn’t necessarily the same as availability – think of a crippled Clariion where one SP fails- write cache goes too – the device is still available but users can be severely impacted.
Chris
Chris,
you are right but the line does get ever blurrier don’t you think? The old distinctions are becoming less important and mid-range is catching up with the high-end; I mean there is less and less you can do to improve the availability of the high-end so inevitably the gap will ever close. Once you’ve got 100% availability, there’s no-where else to go.
Nothing like clinging to outdated terms and ideas..
For example I have a customer who has a number of EMC DMX frames and some NetApp ‘midrange’ as well. I asked him what he thought about the V-MAX announcement and he said “I already have too many old monolithic frames, why would I want to buy a bunch of them and cluster them together?”
I think the old ‘enterprise storage’ market is shrinking and will make like the mainframe- never go away, but the market will be much, much smaller.
Perhaps “External Storage Virtualization that is Enterprise Storage” means that it’s the only virtualization product that uses external (3rd party external to the USP) storage and presents it up to a mainframe with the performance and reliability of typical mainframe Enterprise Storage.
^^^ As if to imply “Enterprise Storage” is Ficon and Escon.
Richard,
if that’s true; the definition of Enterprise Storage is so narrow as to have become pretty much worthless. Enterprise==Mainframe? The market for enterprise storage is even smaller that anyone thought!
I’m not saying thats my definition of Enterprise Storage… I’m an open systems guy. I’m just trying to look at the question and Tony’s answer with an open mind and provided my interpretation.
If you ask a career mainframe guy the same question, “What does Enterprise Storage mean?”… pretty sure they will require Ficon and and ESCON… This question and answer are subject to perception and interpretation.
Ask a mainframe guy anything and you’ll get a one-eyed answer in general. Anything which isn’t a mainframe is a mere toy!!!
As you say, most of the characteristics you would traditionally differentiate on are now a bit moot. Even the cost lines have been blurred (as a quote from a certain vendor proved the other day).
For me, the differentiation between ‘Enterprise’ and ‘Mid-range’ is now one of performance.
As you alude to, you can draw arbitrary lines in the sand as to what is and isn’t enterprise performance, but ultimately, a mid-range is limited by its components as to the speed and amount of requests it can service. “Enterprise Storage” shouldn’t have those limitations.
Oh, and I’d be kicking someone’s ass if I was only getting 99.99% uptime on my enterprise storage 🙂
I have been traveling and took some time off so didn’t see this post until today. First I do want to clarify that I didn’t say that the SVC is not a storage device – I said it was not a storage system because it does not have its own capacity under management.
In any case – there is a difference between Enterprise-class and midrange storage and it is an issue of performance, scalability, reliability and interoperability. There is also different features and functions based on performance and resources in the system. And there is still hardware differences as well that make a difference in all of the above. Whether you agree or disagree in the end doesn’t matter since the market will bear it out one way or another.
I have consulted for companies and have advised them to not put everything on Enterprise-class storage and use midrange for 80% of their data and applications. However, the other 20% still needed to run on Enterprise-class.
Whether you or I think the USP V is or isn’t unique is also academic. It is an issue of value. While you may see little or no value in something doesn’t mean that others agree.
Jon,
TBH I was debating whether Enterprise Class should actually be 5 or 6 nines; I erred on 4 nines!
All storage has limitations; even Enterprise class storage. There are cases where traditionally mid-range arrays may well out perform an enterprise array. What was Enterprise-class performance five years ago would probably now be considered sub mid-range in many cases.
Tony,
sorry for mis-representing you with regards to SVC; I still think IBM will happily compare SVC with USP-V and happily state that they play in the same ball-park.
Yes, there are still differences between Enterprise and non-Enterprise but those differences are becoming less and less every six months or so; traditional mid-range vendors and arrays are snapping at the heels of ‘Enterprise Storage’. And to be honest, I am not any clearer as to what Enterprise storage is, apart from considerably more expensive.
What are the characteristics of the work-loads which need to run on Enterprise Storage? I’m not sure that in most cases it is performance and actually I’m not convinced it’s reliability either.
There is one area where the Enterprise-class arrays do currently win out; that is in the area of array-to-array replication and especially asynch over distance. How long will this area of superiority continue? Who knows!
Yes, Enterprise Storage will still exist as a class of storage for years to come but like the mainframe it so resembles; it will become less relevant to most customers.
Sorry for being later to the show, some random thoughts on requirements re Enterprise Storage :-
* No planned event requires downtime
* No planned event requires functional degradation
* No planned event requires performance degradation
(Planned events are both s/ware & h/ware)
* For an unplanned event the system degrades in predicable & deterministic fashion
* Failure & issue isolation with system
* Fully automated dial-home and remote light’s out support
* Quality of service properties through entire technology stack
* Full abstraction of the exposed capacity from the method used for it’s persistence
* N-way replication & duplication without functional or performance degradation
* Self performance optimising (including sub LUN)
* Full support of multiple disk tiers & types
* No functionality restricted to specific disk tiers or types
* Ability to make multiple parallel / concurrent logical configuration changes without performance or functional degradation
* Guaranteed n-1 & n+1 interop and support
* Guaranteed 7+yr spare parts & software support/bug & security fixes
* FIPS compliant option
* Built in secure erase of disk spindles
* 128+ FC & FCoE interfaces
* Ability to re-configure and migrate storage without host reconfig or interruption
* No requirements for any specific host software for normal operations
* Full built-in reporting on configuration and usage of any licensed technologies
* Ability to support >150k logical disk devices
* Extensible physical enclosure environment
* Full mngt, configuration & reporting s/ware as standard with each system
* Fully active pathing, with no performance impact as to the paths used
* Internal electrical power resilience & controlled shutdown & deterministic restart
Must admit I to am increasingly struggling to differentiate ‘enterprise storage’ & ‘modular’ – often the only real difference appearing to be the cost & hype/fear.
Default position for me is NAS, with requirements needed to justify to Modular and finally to Enterprise.
Will be interested to see how model changes, and the only reason for tiers is € cost saving. If a single platform can be developed that truly allows multiple cost & service quality points then disk product tier conversation changes.
But frankly sick of paying such vast sums for poor quality software 🙁