VFD4 Prep: StorMagic

StorMagic make a Remote Office/Branch Office (ROBO) storage product called SvSAN that provides an iSCSI virtual SAN appliance (VSA) as a Linux virtual machine. It works as a 2-node synchronously mirrored cluster of VMs that provide storage services.

There are some excellent blogs on the details, such as this one by Ray Lucchesi, and this one by Arjan Timmerman.

Focus

Regular readers will be familiar with my love of focus, and companies that really know their target market(s). StorMagic looks like another company with a firm focus on a specific problem and how best to solve it: companies with lots of branch offices (think retail, banks, logistics) that need local apps and data. But they also need simplicity, because they don’t have skilled IT staff at all those ROBO sites (nor should they) and don’t want a large team of staff centrally managing them all. They need scale of management, and ease-of-operations.

Which is hard.

Good Design is Hard

It takes a lot of work to make something easy to use. You have to design the thing so that it is more than fault tolerant, so that it is fault avoiding. The excellent book The Design of Everyday Things by Donald A. Norman describes the concept of affordance: a handle on a door affords pulling, while a plate affords pushing. Many doors have a handle on both sides, but only open one way. How many times have you done something like this with such a door?

This door is actually designed correctly, but many are not.

Once you’ve read this book, you’ll start seeing poorly designed things everywhere. A good tip is that if something has a sign on it to explain how it works, it’s not well designed. There are ways to provide subtle clues to people’s intuition about how something should work, based on their learned experience.

Such as: why is the StorMagic website black text on a yellow background? The whole time I’m there, my brain is going “Danger! Danger, Will Robinson!” because warning signs are usually black text with a yellow background.

SMB Market

StorMagic’s history is the SMB market, which is terribly under-served, in my opinion. This market is hard because they don’t have a lot of money in general, let alone money to spend on IT. It’s an enabler for them, not at all core business, so it has to pull its weight. The price is going to be lower than bigger companies can afford, so you need to have lots of customers to make good money.

That’s good, because you don’t end up depending on a single large customer (which is risky), but it’s a harder business to manage because of the operational scale. You have lots of ‘endpoints’ to manage.

Sound familiar?

That why the companies that serve this market that I like are those that make things simple for both their customers and themselves. If it’s easy to avoid mistakes, you get less support calls. If it’s easy to upgrade systems, it’s easier to sell and support the upgrades. If the systems are fault tolerant and easy to fix, it’s easier to support your many customers. This keeps your operational overheads low, so margins can be spend invested in R&D, marketing, and sales instead of support, which is more of a pure expense than an investment.

And that’s how you make money in a lower-margin business. Ideally you do this in a higher-margin business, too, but in my experience high margin customers tend to be high maintenance customers. I’ve seen what some Account Managers have had to put up with in order to put that second story on their yacht, and they definitely earned the money, though sometimes I wonder how they live with themselves. Still, they can afford plenty of therapy. :)

StorMagic have clearly thought about this, hence features like online target migration (so you can move storage around easily) and VSA restore, where you just drop in a replacement server if one dies, and point the new VSA at the still running node to get things back to full redundancy. The software just handles all the config setup steps, because why shouldn’t it?

Centralised monitoring and management is obviously needed for managing a distributed branch office configuration, and stretched clusters gives you geographic site resilience, because at scale, things will break regularly.

More at VFD4

I look forward to hearing more from StorMagic about their technology and their business. It’s a shame I’m not going to have as much time as I’d like to dig into the background videos from SFD6, but please do go and check them out yourself.