Scality reimagines storage as art

by Robin Harris on Wednesday, 11 January, 2017

The fine folks at Scality send out a new year book of photos and – of course – promos. This year caught my attention because, as a fan of modern art, especially those with Cadillacs, they gen’d up a photo of a disk drive displayed like one of the famous Cadillac Ranch cars.

Here’s an still from a video when the Ranch was fairly new.

Here’s the photo that Scality put out:

The StorageMojo take
Drive vendors might not too happy to see their products treated so cavalierly, but it only reflects their iconic status in computing for the last 6 decades. And I like creative marketing!

Courteous comments welcome, of course.


Violin’s bankruptcy

by Robin Harris on Friday, 6 January, 2017

Violin Memory, one of the early entrants with an all-flash array, filed for bankruptcy last month. The company continues to operate under Chapter 11, but this is a sad outcome for a pioneer.

So much for first mover advantage
When I first met with Violin, the original team had a great idea and not much traction. The architecture connected the flash storage nodes with a mesh network to ensure dual-channel access and high-performance, with system visibility down to the individual flash dies.

This was at a time – 2008 – when most folks (other than Fusion-io) were looking to emulate disk drives with flash in a can. This seemed suboptimal because, after all, if we had invented flash in 1957 instead of disks, is that what flash drives would have looked like 50 years later? Of course not.

I thought then, and still do, that this was an innovative architecture. But innovation can be hard.

Architecture isn’t everything
I was critical of folks, like Pure, who put SATA drives behind a couple of controllers, and called it good. But this was a time to market strategy, so they could focus on software, not hardware. And software is a lot cheaper and quicker to build.

Which is, ultimately, what cost Violin their early lead. By concentrating on hardware, Violin found it hard to compete with feature-rich software implementations, despite Violin’s excellent and consistent low latency. Buyers were dazzled by IOPS, not latency.

The StorageMojo take
The race is not always to the swift, nor riches to the wise. By starting with software, other companies built an early lead, and now have the money and time to optimize hardware for flash.

It looks like a new race is starting, with NVRAM as the instigator. I’ll try to remember the lessons of Violin’s path.

Courteous comments welcome, of course. Happy new year to all. As for me, this new year has gotten off to an excellent start! Also, I’ve done work for Violin.

{ 1 comment }

Thunder Mountain

by Robin Harris on Monday, 26 December, 2016

StorageMojo’s hike blogging has been on hiatus for a few months, due to a personal issue. But no worries! If all goes according to plan I will be better than new by the end of January.

I’m more than ready! I’ve had to skip too many industry events, such as Flash Memory Summit and the upcoming CES. But I expect to attend the Usenix File and Storage Technology (FAST) event in February.

Starting the day before Christmas, and continuing through much of Christmas day, it was snowing. Yes, Virginia, there is snow in Arizona. Parts of northern Arizona could easily be mistaken for Idaho or Montana – in climate as well as topography.

But after the storms, the sun comes out, and I hiked out onto my deck to take this picture of Thunder Mountain, aka Capitol Butte.

Click to enlarge.

The StorageMojo take
Despite my difficulties, this holiday is a more than usually joyous one for me. I hope you can say the same.

Happy New Year, to one and all.

Courteous comments welcome, of course.


Purpose built backup appliances: cloud collateral damage

by Robin Harris on Thursday, 22 December, 2016

It makes sense that the WW purpose-built backup appliance would be suffering. Cloud-based data gets IaaS provider DR, while cloud backup software handles day-to-day backup, and modern object storage systems optimize archiving.

Back in April of 2012, IDC produced a PBBA market analysis that predicted that the PBBA market would be $5.9 billion by the end of 2016.

Well, here we are at the end of 2016, and guess what? The PBBA market is well short of that.

How short?
Eyeballing the current stats, it looks like the IDC stats for 2016 will come in at about $3.5B, 40% below the 2012 IDC prediction.

Unlike other mature markets, it also appears that leader EMC’s share, courtesy of the Data Domain acquisition, is also drifting down from its ≈65% share in 2012 to a still dominant ≈60% in 2016. I would have expected the incumbent to pick up share as other vendors, starved for revenue, retreated.

The StorageMojo take
The cloud’s collateral damage to the legacy IT vendors continues to spread. A few billion here and a few billion there, and pretty soon you’re talking real money.

And yet, as I recently noted on ZDNet, users still overspend on IT infrastructure – both physical and virtual – by many 10s of billions of dollars every year. With the growing transparency of IT costs – check out this CloudPhysics quarterly report, the historical and organizational forces that favor overspending are weakening.

Vendors who can show advantages in cost and efficiency are best positioned for continued slow-motion clash between on-prem and cloud infrastructure.

Courteous comments welcome, of course. Sure, the IDC forecast could have been based on silly assumptions, but I don’t think so.

{ 1 comment }

The myth of video anonymity

by Robin Harris on Tuesday, 13 December, 2016

Artificial Intelligence has achieved breakthroughs that directly affect documentary and investigative reporting, or any video where participants need anonymity. Thanks to advances in artificial intelligence (AI), standard methods of cloaking identities through pixelation and audio adjustment are much less effective than they were even five years ago.

Lives may be at stake
AI, maybe you’ve heard of it. A recent application of AI is in an area called Image Super Resolution or ISR, which uses AI to scale up lossy or pixelated video to determine the most likely content that the video sensor caught. Hollywood uses a fake version all the time: the getaway car is caught on a grainy surveillance video and the hero barks enhance and, miraculously, the license number appears.

The reality is a bit different and scarier. A pixellated face and a disguised voice can be reverse engineered to determine an individual’s identity, leaving them open to reprisals.

A technology called convolutional neural networks is the tool of choice. They’re very good at sorting through possibilities quickly, and they never tire.

The recent paper Amortised Map Inference For Image Super-Resolution, machine learning researchers Casper Kaae Sonderby, Jose Caballero, Lucas Theis, Wenzhe Shi & Ferenc Huszar, all of the Twitter Cortex in London, describe a new approach to enhancing images.

Results are amazing. Starting with this image:

They get this:

The method works better, today, on natural features, such as faces or a tree. It can’t single one person out of 6 billion. But if a repressive regime knows who the top 500 activists are, chances are good that an AI/ISR bot will identify them, despite pixellation.

The research on ISR is vast and growing, and it wouldn’t surprise me if some of it were funded by security agencies here and abroad. There is similar work being done to undo audio cloaking as well.

What to do?
Clearly, pixellating faces no longer protects vulnerable people. I believe the most likely solution for video is to replace a person’s face with a generic face and then pixellate that. For voice the audio might have to be read by another person and then disguised, or perhaps converted to text and read by a computer voice.

But I’ll leave the solutions engineering to others.

The StorageMojo take
Most of the angst around AI is focused on human-level – or beyond – intelligence – HAL 9000, Skynet – rather than highly specialized AI. But the latter are easier to develop and can deliver astounding results in limited domains.

Today about half of all American’s faces are in a police database. Videos are going to have to smudge or replace faces before pixelation to preserve anonymity.

At a high level this is simply using AI to add value to the imaging systems we already use. The payback scales with our cameras and storage. Expect many more analogous systems.

It’s the applications of these systems, and their unforeseen consequences, that we’ll want to have a say in. For example, what if the enhanced image looks like you – and you weren’t there?

Courteous comments welcome, of course.


Nantero raises $21 million and that’s good

by Robin Harris on Tuesday, 13 December, 2016

Nantero raised a $21 million round from investors. The company is one of StorageMojo’s favorite NVRAM vendors, because carbon nanotubes.

I also like the fact that their process can use existing fabs, even fully depreciated ones, to build high-density vertical NVRAM. The business model takes after ARM rather than Intel, which means they can harness the R&D budgets of multiple firms.

I usually skip the boilerplate quotes in press releases, but this one from the round’s lead, David Poltack, Managing Director, Globespan Capital Partners, is on the money:

Nantero has multiple industry-leading customers who would like to receive NRAM even sooner. The fact that several of these customers, as well as key partners in the ecosystem, have decided to also invest in Nantero is a strong sign of confidence given how well they know Nantero and its product from years of working together.

The StorageMojo take
NVRAM is the Next Big Thing in storage. StorageMojo expects that it will have a much greater impact of systems and storage than NAND flash has had over the past decade.

Why? Flash added a new level to the storage pyramid and made IOPS cheap. NVRAM will rearchitect the storage pyramid, possibly replacing RAM, flash and, depending on cost, disk, with a single flat address space. Since CPUs aren’t getting faster, learning how to best integrate NVRAM will be the next big battle on the performance front.

Courteous comments welcome, of course.


Cloud market heats up

November 18, 2016

In 2014 Gartner estimated that Amazon Web Services had 5x the utilized compute capacity of the rest of the cloud providers. There’s a couple of qualifiers there – utilized, compute – but as a rough guess, it looked like AWS had around an 80% market share. But no more. In a recent report Synergy Research […]

1 comment Read the full article →

When is a feature a bug?

November 17, 2016

Ten years ago in Enterprise IT: the elephant’s graveyard I wrote about the upmarket trap: Engineering and marketing find it easy to justify fun new technology since a 10% goodness increase on a $500,000 machine is worth $50,000, while on a $1,000 machine it is worth $100 only if the customer is knowledgeable enough to […]

1 comment Read the full article →

Frisky Gen-Z’s to battle boomer Intel

October 24, 2016

The internalization of storage is spawning another war – this time in memory interconnects. From Anandtech: This week sees the launch of the Gen-Z Consortium, featuring names such as ARM, Huawei, IBM, Mellanox, Micron, Samsung, SK Hynix and Xilinx, with the purpose of designing a new memory semantic fabric that focuses on the utilization of […]

0 comments Read the full article →

ClearSky: object storage at enterprise block speed

October 17, 2016

Can object storage ever be as fast as block storage? It turns out the answer is yes. And we already know how to do it. I was speaking to the CTO of ClearSky Data, Laz Vekiarides, about their block storage system for enterprise applications. They offer . . . a Global Storage Network that manages […]

4 comments Read the full article →

Everspin’s MRAM IPO

October 10, 2016

Everspin has filed for their IPO. They’re looking to raise $40 million from the public market. They’ve been shipping product for over 10 years, so this is a real company, not a dream and a slide deck. Why MRAM? Everspin’s Magnetic RAM has a number of advantages over flash and DRAM: DDR RAM write latency […]

2 comments Read the full article →

Is 3D XPoint in trouble?

October 6, 2016

The Register’s Chris Mellor and SemiAccurate’s Charlie Demerjian are throwing shade on Intel’s claims for 3D XPoint. While it’s great fun to tweak the giants of tech – as I often do – I think they are likely wrong in their interpretation of the data backing their arguments. As Mr. Mellor wrote: In contrast to […]

0 comments Read the full article →

Nutanix IPO: the big score

October 5, 2016

Nutanix – NTNX – has started off with a bang: opening at $16 a share and quickly rising to almost $30. It’s trading at $36 as I write. Now for the hard part Everyone is no doubt counting how much they’ve made on that spectacular beginning. But there’s a six month lock-in, meaning insiders have […]

1 comment Read the full article →

Dell vets: buff up your resumes this weekend

September 9, 2016

Now that Dell has completed the EMC acquisition, you are in for a rude awakening. While Dell may own EMC, EMC owns you. Richard Egan, one of the founders os EMC, fostered an exceptionally aggressive sales culture. The company liked to hire guys from blue collar families who’d played football in college, and then set […]

0 comments Read the full article →

Artisanal science doesn’t scale

September 8, 2016

Big data will overwhelm artisanal science. That’s what I conclude from a recent paper that lays out the stark statistics: Science is a growing system, exhibiting 4% annual growth in publications and 1.8% annual growth in the number of references per publication. Together these growth factors correspond to a 12-year doubling period in the total […]

1 comment Read the full article →