Profile cover photo
Profile photo
Smarter, Faster, Cheaper Networked Storage
Smarter, Faster, Cheaper Networked Storage

Coraid's posts

Post has attachment
Check out the new #Coraid HBA for VMware:

Post has attachment
From Shugart to Seagate
Alan Shugart and his business partner Finis Conner sat across from John Young, the then-CEO of Hewlett-Packard. The meeting was not going well.

“Al,” Young asked, “why should we pay half a million dollars for 25% of a company that is only an idea in the minds of you and Finis?”

“Perhaps you shouldn’t,” Shugart shot back.

Young and his fellow VCs from Page Mills Group declined to invest.

But that didn’t stop Shugart, it just slowed him down for a day. Shugart’s life is a study in ups and downs, one I think you’ll find inspiring.

Read the rest on EtherDrive Blog @

Post has attachment
The Power of Technology Layers
The core•aid SR704 doesn’t run on the operating system. It runs in the operating system.

This makes the SR704 wickedly fast and incredibly stable, but understanding why is overwhelming without first understanding the layers involved. Heck, even understanding how your laptop works is mind boggling when viewed from the top. How on earth can one person build such a complicated system?

The trick is dividing the complex into simple layers, each with an interface, and the system becomes manageable. Layering.

There are examples of layering everywhere. For example, the internet began as one massive complexity until DARPA engineers realized there were two problems to solve: the internetworking problem (moving data between networks) and the transport problem (ensuring bytes made it to the other end in the correct order and unduplicated).

This became TCP/IP. Two tin cans and a string.

Read the rest of the article @

#computing #etherdrive #technology

Post has attachment
Production + Research and Development is Key to Innovation
Thomas Edison was working on a new telegraph transmitter when suddenly a speed regulator broke. As the paper tape flew through the device, the noises it produced were eerily human. This accident started a chain-reaction of experiments that led to the world’s first phonograph.

Accidents are the root of many technical innovations, which is why we need to leave room for them in our organizations.

Research and development must be closely tied to engineering. If it isn’t—if developers have tunnel vision on a planned product release—think of all the missed serendipitous events that could have led to great products.

When Robert Noyce and Gordon Moore left Farchild Semiconductor to start Intel, they made a point to house research and development with production manufacturing.

Moore, former head of Fairchild’s R&D, knew firsthand the frustrations that arise from separation. So he organized Intel to have both departments in one unit.

This model allowed for a wide variety of technology innovations. For storage, it led to one particular change that would forever alter how we think about data.

The Fabrication-Line Accident
In 1971 Intel engineer Dov Frohman scratched his head as he tried to puzzle through what went wrong. In front of him were a bunch of chips fresh off the fab line, every single one of them non-functional.

The mystery continued: each one went wrong in a different way. They’d all had the same data written to them, but their outputs seemed to be random.

Frohman found it even stranger that if he put the quirky wafers in sunlight, their outputs changed permanently.

He cut one open and stuck it under an electron microscope. He quickly realized what went wrong.

The chips had too many layers. Manufacturing must have run the wafers though the fab line twice!

Read the rest of this article on EtherDrive blog @

#innovation   #happyacidents   #engineering   #researchanddevelopment   #intel  

Post has attachment
Is 2016 the Year of All-Flash?
Have you ever flipped open a newspaper and read a “fact” that made you think “that can’t be right”?

I had one of those moments while reading the Boston Business Journal: “We believe that 2016 is the year of all-flash for primary storage as we reached a point where it can cost less to store data entirely on flash rather than performance hard drives.”

Solid-state storage costs less than hard-disk drives? My gut feeling is that this handy projection is blatantly false, but I’m not one to bet on gut feelings. I like facts.

So this post is a small thought experiment with one goal: find out a ballpark price difference between HDD and SSD storage.

(Before I launch into my experiment, a quick warning for readers who want to verify the quote: you’re gonna be slogging through some serious startup speak. This guy was inventive. Instead of Steve Jobs, he compared himself to Elon Musk. Groan.)

Read the rest of this post on EtherDrive Blog @

#allflash   #ssd   #hdd   #datastorage  

Post has attachment
Coraid’s Command Catastrophe
A middle manager in Coraid’s California offices, let’s call him Goandbeckon, once sent an epic email rant about the command names for the Coraid SR RAID array. Goandbeckon railed about one specific command, digitally shouting that it was an “embarrassment” to the company.

Goandbeckon’s complaint? The command to display disk drives was named… disks.

So Coraid radically renamed commands in the last few releases of the SRX before it went under (as I wrote last week, this became the foundation for core•aid SR704).

As a result, a whole lot of users and support folks were immediately frustrated. After ten years of using a particular command set, people’s fingers started working on auto pilot. New command names threw an unnecessary wrench in their day-to-day routines.

When SouthSuite built the core•aid SR704, I went back to the original command set. Here’s why, along with a quick run down of the SR704’s command simplicity.

Read the rest of this article on EtherDrive Blog @

#commands #coraid #etherdrive

Post has attachment
How Not to Impress a Tech Audience
During a weekend Netflix binge, my best friend turned to me and said, “If these guys mention Steve Jobs one more time, I’m turning this off.” 60 seconds later, we were hunting for something else to watch.

I checked the timestamp: we only made it 8 minutes into Print the Legend, a documentary about 3D printing. In that time, I learned nothing about the technology or the people making it, but I learned a lot about how technology companies should think about communication.

In the first 8 minutes Print the Legend referenced Steve Jobs four times, including showing a clip from the Apple iPhone keynote. 3D printing aficionados used the term “revolutionary” 6 times and mentioned “changing the world” 3 times.

That’s a bull$@?! per minute rate of 1.6 and my buddy and I couldn’t sit through more.

Yes, this was a documentary, but the mistakes made were ones you see all too frequently from tech startups.

We essentially watched a poorly constructed introduction. I say “poorly” for two reasons: the documentarians didn’t know their audience and, if their story had any substance, they waited so long to tell me about it that I lost interest.

At some point, you will introduce your technology to a market. Don’t make these two fatal errors.

Read the rest on the EtherDrive blog @

#startup #tech #communication #etherdrive

core•aid Shipping SR704
After wrapping up a brief testing period, I’m proud to announce that SouthSuite is shipping the first software release of core•aid SR704.

The SR704 is a software version of what used to be called the SR and then the SRX, the AoE technology that sold almost a $100,000,000 worth of hardware systems.

The SouthSuiters and I took the most stable version of the SRX core and cleaned up the clutter and clunk that had built up slowly over the years (mostly because of “not quite in touch” product-management requirements from the SRX’s previous owners).

I’ve made it easier to update, easier to support, and easier to think about.

Central to the idea of easy is the AoE target. In a SR704 storage appliance, the target appears to initiators on the network as a simple disk drive. That “disk” can be either a single drive or a redundant array of cheap disks formed into a single target. Maybe they should have named it RACD instead of RAID.

The core•aid SR704 release has left us with a “perma-grin” for a number of reasons.

Read the rest of this article on EtherDrive Blog @

#coraid #AoE #datastorage #etherdrive

Post has attachment
Learning from History: The RAMAC 350
Learning the history of a technology gives you a story arc that lets you see where that technology is heading. The invention of the disk drive, for example, sheds light on the current evolution of storage technology.

In 1952, IBM created a lab in San Jose, CA, under Rey Johnson’s leadership. One of the lab’s missions was finding a way to inexpensively automate something called a “tub file.” Never heard of it? No wonder—it doesn’t exist any more.

Automating Data
In its heyday, the tub file was a collection of large trays that contained punch cards in some predetermined order. It allowed relatively fast, random access to data, supporting the workflow of companies in the pre-computer era. Improving this process up was important and it became a prime candidate for automation. (You can see a video of them over on EtherDrive’s Facebook page.)

In the 1950s automation was a big topic of discussion in the business world.

During the previous 50 years, specialized machine tools had reduced the cost of production a great deal, but people still had to shuffle data around mostly by hand—they had to find the appropriate stack of punch cards and load them into a hopper.

“Automation” meant inventing machines to move this work from machine tool to machine tool, and IBM’s customers were asking for automation to replace the time consuming manual operations involved with the tub files.

The storage technology of the day consisted of punch cards—invented in the 1880s—and the more recent storage mediums, tapes and magnetic drums. They each had their problems: cards and tapes were sequential and best used for batch operations. Drums took up too much room and held too little data.

One device that came to the group’s attention was the “Notched Disk Array” invented by Jacob Rabinow in his work as a consultant to a new computer development group within the US National Bureau of Standards.

This device accelerated and stopped a disk coated with magnetic material over a read/write head that merely pivoted from disk to disk, much like a Wurlitzer jukebox moved 45 rpm records (for more on jukeboxes and magnetic recording, check out Charlie’s post, ”Inventing the Hard Drive: Jukeboxes and Tombstones”).

Although patented, Rabinow’s device was only a lab experiment and wasn’t practical for IBM’s customers, but it did spark the idea in San Jose’s collective minds of how to put the data on disks coated with iron oxide.

Read the rest @

#RAMAC350 #IBM #HDD #SSD #storage #computerhistory #IBM #EtherDrive
Wait while more posts are being loaded