Why the USS NetApp is a doomed ship

This has been an ongoing train wreck happening for well over a decade.

It's sad as they were once a great place but somehow their products haven't evolved and they

never fully embraced the advent of flash being anything more than a jammed in disk tier or cache. Sometimes they stuff it into their Lsi arrays and practically give them away but they lack layer featured that were astutely called table stakes in the article

They are sadly unable to acquire companies without rapidly killing them or their core IP somehow. Coral was great about spinfs but it somehow died and cmode seems to leave customers angry for one reason or another.. Customers often don't know why they have to move to cmode but they are being forced to and still don't get it. Cmode is more about ntap than the customer..

I feel bad for them but they are ending up irrelevant and not engaged much anymore.. They still have the best Transactional NFS system on the planet for disk but that is the last piece that keeps them alive at this point.

I don't seek out people to have negative conversations about ntap but it seems people seek me out to tell me how unhappy they are with the products that ntap produces, their support and the entire company. I remember when they were loved but customers seem to have strong disdain for them now.. I believe their netpromoter score is below 20 (17) and declining.. I sincerely wish they find a way to turn this around but it's far too late and they have had so many chances at this point.

The auther is right - and it's a shame

Opinion:

I remember sitting round a table well over a decade ago now explaining to my colleagues about the key differentiators that NetApp storage had over the competition. It was such an easy argument to win on paper and I made it core to my career, whether in enterprise or consultancy.

Here's what killed it off for me about three years ago:

1 - The end user experience

The solution was always expensive but it could deliver such a wide feature set and still manage to perform reasonably up to and including the x1xx range (61xx 31xx). When the x2xx range were delivered, in spite of the speeds and feeds, the solution defied logic and appeared to cease providing the linear performance increase that you would expect with a next generation tech refresh. It was as if the underlying OS was incapable of exploiting changes in CPU architecture. I've not compiled stats to support this - it's only anecdotal stuff. In fact, in one case I was sure that I was seeing better throughput from an older generation mid range controller than I was a newer one.

All the while, user expectation for performance followed general trends and the chasm between what I could deliver and what people wanted widened. Support quality took a dive (sorry but true) and complaints from customers meant repeated churning of that most cumbersome and hated troubleshooting tool, the perfstat. Considering the autosupport mechanism is already in place, you would have thought that NetApp would have made the investment to integrate perfstat in to that reporting loop rather than forcing people to jump through that kak handed RSH hoop.

2 - C-Something or other

During the extended lifetime of the x1xx hardware release a couple of things happened. NetApp took their foot off the gas for 7 Mode development and started throwing everything in to C-DOT / C-Mode / Cluster Mode whatever you wanted to call that decade old project. Very slowly, 7 mode began to deflate as NetApp started to blow all the hot air in to version that went live before it had matching features. At the same time, the need to do things at the array level lessened as people began to solve problems higher on up the stack. The market doesn't stand still - and if you take too long to deliver your new product, you have to consider whether it's still going to be relevant.

3 - Competition arrived

Others just started doing it better. Whilst the closed feedback loop of CDOT was happening all over the folks at NetApp, real innovation began springing up elsewhere. You cannot tell me that the Nimble isn't a genius product when you can push hundreds of MB read / write on just a handful of SATA disks. You can't tell me that Nutanix obliterates the needless complexity of monolithic storage + compute in one swipe. And even if you're looking for a like for like traditional product, v7000 delivers many of the features and stands up to performance demands - and I've only every seen the transition toward v7000, never back the other way.

As all this went on, NetApp failed to innovate to the level of their competitors and they failed to acquire the right stuff. They appear to be trapped in bubble of C-DOT and Flexpod and I suspect they've turned up the marketing volume in an attempt to maintain their profile and drown out the competition. It's going to be interesting to see what happens to their market share.

That's how I see it from my corner of the world and my own experiences. It doesn't mean that the experience is like that universally but, as a previous poster said, I find very few people in the industry praising them these days. Change is the only constant.

Maybe some facts?

There will be other readers I suspect who will look at articles like this where there is intense - polemic - discussion without any detail and feel somewhat disappointed..

What is Netapp failing to do? What exactly is the future compared to the present? What new features must be included and how difficult is that to do? How many people actually need the features that Trevor is thinking about (but not being specific) and what evidence is there to support all this?

I know journalism requires people to make noise even when there is absolutely no need for it but IT is riddled with constant new terminology - new anodyne, abstract words that imply innovation but usually are just rehashing existing ideas. Reg readers are more likely than average to be irritated by generalities and expect to see more facts. Without knowing what the real criticism is, it is difficult to judge whether the whole premise of the article is accurate or just hot air.

Really conservative for this era..

The article is true for most of it's part. Actually the conservative stance the company takes is puzzling. With about 5B in cash and reasonable debt levels, they had umpteen choices 2-3 years ago but chose none. They could have spun-in an R&D division with fresh talent load from top companies (or) could have bought a startup with a reasonable customer base (or) internally fired up innovation in house to re-architect and re-write Ontap like system for modern data loads/patterns. None of these happened which is appalling. While EMC is no great internal innovator they buy companies to plug the gap which has helped them stay away from really bad y/y declines. Tonnes of money have been thrown into share buy backs (exactly not sure for what!), as most people I know who do a good job have little or no RSUs - might be to just serve top guns and keep the share price high. The sad part is "It's hard to wake up some one who isn't sleeping".