Scale Computing is a tiny fish in a small pond. Fancy its chances?

Scale Computing is one of 13 suppliers attacking the hyper-converged infrastructure market. Not all will survive. What has it got that makes it distinctive and gives it the potential for success? Scale’s difference is based on its SMB customer approach, meaning low-cost and simplified admin, and cleaned up IO stack. This, it …

[Trevor_Pott]

Review coming

I have had a chance to review Scale's new hybrid nodes and am just finalizing a review. These nodes are going into a customer deployment where we are collapsing 8 VMware nodes and a couple dozen physical servers into 4 Scale nodes. So far, the experiment has been a rousing success.

I also run my own company on 3 (admittedly aging) HC1000 nodes. After ~1 year worth of running workloads on it, I don't have any complaints. It has survived power outages and dead drives, network weirdness and other things. It does its job.

Scale isn't as feature rich as VMware. It never will be. For the target market, however, that simply doesn't matter. Small businesses make up 98.2% of employer businesses in Canada. Enterprises (500 seats and up) make up just 0.1% of employer businesses. In the US the percentage is a little different, but the rest of the world looks very much the same.

So Scale is one of a handful of vendors that serve the 99.9%. They don't meet every niche. They don't cover every possibility...but they solve most problems for a price that SMBs can actually pay. Perhaps more importantly, they listen. That's something few vendors do.

I can't speak to other geos, but in Canada Scale is doing quite well. Some of their competitors (Yottabyte, for example) are making headway here too. Even Tintri - who normally focuses on bigger fish - is starting to find its feet in the mass market.

The old guard of server vendors are no longer the only ones. Marketing is starting to make an impression even amongst SMBs. More importantly, once you give a lot of these new vendors a try, it's hard to go back. They make ease of use kind of their "thing".

All the above being said, take precautions. Do a POC. Don't buy blindly. Test the products first to make sure you understand all the differences in UI, that you know how move workloads across, and that you are comfortable with the featureset provided.

You needn't fear the new crop of vendors as much as many doomers in purchasing clinging to their junkets, hookers and blow will say. But you do need to keep your eyes wide open; new vendors are a new mix of features and a new support commitment. Most will be happy, but better safe than sorry.

[Art Jannicelli]

Back in the late winter of 2013 I had the displeasure of dealing with a Scale disaster, the client lost a single disk in the array. No problem, they got a replacement disk and inserted it. The whole array locked up and would not respond on a Friday night after 5pm. We spent the entire weekend on DR, Scale support tried a different disk, sending us a whole new node, nothing would bring the array back online.

Long story short we had to take the disk to an expensive drive rescue place to have the platters relocated to a new disk. Scale support was then able to get the array back online without any data loss (other than what we had to reintegrate from DR). Scale never came back with a good answer on why/how it happened. My client migrated off ASAP and it is now used as dumb storage.

After that horrible experience with support and their hardware I'd think hard about using them for production data.

[Trevor_Pott]

I'd be very, very curious to find out who that client might be. Considering the number of scale installs I know about in the wild - and the number of them that have handled dead disks, node replacements and more - I would love to sit down and have a detailed discussion about the event.

If it was legitimately an issue with Scale's support or product, I'll pick up the phone and have a discussion with their CEO and we'll see what (if anything) has changed to ensure this doesn't happen to others. If it was an edge case phenomenon it's worth knowing what the exact details are so that I can write something up for Scale customers so we all know what to avoid. And then work with their devs to make sure that it can be reproduced and then prevented.

Please ping me using my contact page, if you can possibly share the info or - better yet - put me in contact with the client. These are pretty serious claims, and are worth investigating. Especially with so many Scale customers out there, and the growth that company is seeing.

[Art Jannicelli]

Re: Xpenology

Thank you Trevor. He reached out to Scale and determined my client was running a different pre-scribe OS and experienced a known issue of that particular OS that is not applicable to this product.

[Huw D]

A friend of mine has implemented HC3 at a company in Nova Scotia and there's no complaints at all. It's going to be one of the options I'll be looking at for a client of mine next time they do a server refresh.

Plus I've got a lot of time for the lads at Scale and I probably owe Jason Collier a beer or two. ;)

[Trevor_Pott]

Huw! How ya been, ya crazy son of a gun? Long time, no flaming beverages!

[Lost_Signal]

Couple Thoughts...

1. I"m curious who are they getting SLC flash? High endurance flash today (Like an Intel S3700 that has an impressive 10DWPD) is eMLC. I know Micron and some will custom make whatever you want, I'm just curious for endurance reasons why you would go SLC when eMLC can do quite a lot.

2. Essentials Plus is ~6K for the capex and ~1K a year for support. Saying the "vTax" is significant is a bit of a stretch for

3. Having seen a VSA use 8 Core's (and half of a HCI host) I agree that the VSA architecture has problems scaling (Either out or down). Kernel implemented storage makes the most sense for HCI for CPU and memory reasons.

4. I have a lot of respect for the Scale guys. They are not following Go 4 billion or bust plans of other startups in this space, but a methodical customer focused route. I wish them well, as I think this would be a better model for taking care of customers than many of the existing startup models.

[sempertyrannis]

Re: Couple Thoughts...

Hey @Lost_Signal - we introduced eMLC this week in our new 1150 platform. keep your eyes open for forthcoming reviews

[Trevor_Pott]

Re: Couple Thoughts...

2. $6000 up front with $1000 per year is a lot of fucking money to many business, John. As you well know, but have never actually cared. What's more, you're completely ignoring the shitshow that is going from an EP license to "more than 6 cores", which is roughly as much fun as burning off your own genitals with acid and about as messy.

3) *yawn*. Not everyone's VSA is as horrible as Nutanix's. Maxta, and Simplivity both have great solutions that don't expend your cluster's resources running the cluster's storage. And at least with someone like Maxta you can actually use the damned software on something other than the narrowest slice of the highest end of hardware. VMware's storage team allows for approximately the square root of a gnat's fart worth of the world's hardware to be used, so what the hell's the point?

4) Scale are absolutely "go for billions or bust". They just plan to take a little longer to get there, and build an actual business underneath them in the meantime. Fortunately, they aren't the only ones with this model, and the HCI space now has many decent contenders offering HCI capabilities on a wide variety of hardware , with different price points and - most critically - the ability to scale from one node to many without getting randomly stabbed in the junk by unexpected massive leaps in licensing costs.

Hurray for THE FUTURE!