This makes no sense "...an architecture that pretty much looks identical to the one we know and "love" from NetApp".
NetApp is not a monolithic, architected 'thing'. I can only guess that the author means there are common features.
Barely a day passes without receiving an announcement from some vendor or another. Every one is "revolutionary" and a massive step forward for the industry – or so they keep telling me. It would seem that we are living in a golden age of invention. Yet many conversations with peer end users generally end up with us feeling …
In my humble opinion, as a buyer of storage, there are some interesting innovations, particularly with cloud integration (yes, I know, YMMV, always read the label etc.). But I also recognise some of the woes Storagebod outlines - particularly the excessive VDI focus, which to be honest is rather niche for most of my clients.
Salesforce.com is the ultimate - everything is announced as 'the first', with excited claims of 'transforming' businesses, markets or whatever, even if all that the company has done is caught up with what its competitors were offering a decade or more ago.
That's a bit Apple
My god! Multitasking? That's amazing! What?! You mean I can copy and paste text from here to here? Swoon! Shaped earphones? Whatever will they think of next?!
But most companies do that - advertising any new feature as though it were a breakthrough that will change the world as we know it. A touch screen tablet with a removable keyboard that you can use to run not only business programs but entertainment software as well, MS? I think I need a lie down . . .
going on these days, I suppose the most obvious one is SSDs, but a good storage system is a hell of a lot more than just slapping a bunch of SSDs in it and allocating LUNs or exporting file systems. Finding good ways to best utilize the underlying technology's performance while maintaining high levels of reliability/availability and if your lucky scalability too.
I don't think there is much over hype in the storage industry at all. Now by contrast the networking industry and their software defined bullshit THAT is hype(hype to roughly 99.999% of organizations anyway). Networking has traditionally been a very boring thing, and they are desperately trying to get attention with software defined.
Some folks are trying to capitalize on that and offer "software defined storage" which I view as hype as well. There are some interesting approaches but too many platforms are claiming software defined when it doesn't make any sense.
That said I still find myself a loyal 3PAR customer/advocate 8 years after getting my first system. They continue to give me reasons to stick to them. I told their leadership at Discover that if you asked me two (and for sure three) years ago did I think the 3PAR architecture would make it in the SSD world I would of been very skeptical. What they've managed to pull off without having to do any ground up rewrites or acquisitions in the past 18 months has just blown me away(and there's more to come as always).
I do see interesting things that some storage startups are doing - if I worked for a much bigger org that had more silos of storage laying around I'd be interested in test driving them. As-is, for the most part anything we get has to be capable of being a backbone driving tens to hundreds of millions of revenue each year(over $200M/year is driven by roughly two racks of equipment today). Which for me means I don't want to take the risk and use the startups for that kind of thing - I've learned a lot about storage over the past 10 years or so and perhaps the most important lesson is I've learned to be conservative in the risks taken(and yes I have had my fair share of issues on 3PAR over the years, by no means is it a flawless platform - which just re-iterates my feelings against relying on the startups for such a system).
(not a storage guy specifically I do networking servers, ops stuff etc as well)
He's right about the software-defined bollocks though. It's a bandwagon, and a pathetic, utterly meaningless one thought up by marketing idiots and lapped up by moronic salesmen who don't understand what they're actually selling and are more interested in badgering customers into thinking they are buying the latest and greatest technology which will last them for years.
All hardware devices have software in them. If they didn't they would be lumps of metal and silicon. It's not rocket science to rewrite code to work with different hardware. We've all been doing it for decades. We just didn't call it software-defined.
Give it a few years and they try and tell us that software-defined is old-fashioned, just like they're trying to do with virtualisation. And in the meantime, it's the same stuff underneath, with different pictures on the top.
The low-level stuff is continuing to incrementally innovate. Has been for decades: Drives get more capacity, SSDs likewise. This is component improvements, driven by people who are educated in such esoteric fields as quantum mechanics and magnetic domain modeling.
Above that though, what is there to innovate? You can't turn X bytes of storage into >X bytes through any form of elaborate software misdirection, you can only provide the illusion. You can't recover your X bytes of data if less than X bytes remains after a drive failure, so you always need X+Y for redundancy. The mathematics is unforgiving. You can only try to make the best from the hardware with tiering and management functionality, a field which is pretty much done to death now.
Biting the hand that feeds IT © 1998–2021