Year end thank yous
2016 was awful, but some vendors weren't.
It’s that time of year again. The time of year where I thank vendors for sucking slightly less than their competitors over the previous 12 months. Some do this through product excellence, some through excellence in customer service.
As a rule, I find vendors trying, and so I take an almost absurd amount of pleasure in reminiscing on those that, in one form or another, make the grade. To this end, I will start the year’s festivities with a joke about how to define how “ready” a product is for general consumption.
Google ready: 4 years, 3 PhDs and an army of undergrads later it runs in 30 datacenters on 5 continents, but only two people know how.
Hyperscale ready: with a trained $1M/yr specialist and a support team of 5 it will support 50,000 clients at 99.95% uptime.
Enterprise ready: with a team of 20 certified technicians, 4 middle managers and a vendor engineer it will integrate with our mainframe.
SMB ready: one of the two burnt out husks that were people remembers enough about how it works to restart it.
Trevor Pott ready: issues reduced to one 4:00am phone call per annum.
So without further delay, here is my totally biased list of 2016’s kickass vendors that have made my own personal life easier:
Scale Computing make hyperconverged appliances. Servers (3 or more) with built-in storage that work together to allow virtualized workloads to achieve high availability without needing a SAN. Scale’s solution is built on top of KVM and uses their own custom software for monitoring, self-healing, management and a user interface.
Scale doesn’t expose nearly as many of the hypervisor’s features to the user as rivals like VMware. From one perspective, this means that Scale is nowhere near as feature-rich as VMware. From another, by being built on KVM, they have access to almost every feature VMware doesn’t, and a few besides. What they choose to expose to the end user is where things get blurry.
Scale’s strength is their ease of use. Their product “just works”. It’s more than fast enough to handle my real-world workloads. My customers love it. I love it. My Scale clusters are the only pieces of infrastructure in production use today other than my switches that I can honestly say I don’t have to think about for the vast majority of the year.
Thank you, Scale. You reduced weekly 4am screaming about aging VMware + SAN infrastructure to no screaming at all. Sleep is good.
What can I say about Tintri? Tintri sent me a T-850 early last year. I kicked the crap out of it in the lab, and it stood up well. I then had a customer emergency that required solid, reliable storage now and Tintri graciously allowed me to put their demo unit in place.
Well it’s been a year and a half, and for various reasons to complicated too bother with, the Tintri is still in place. For a year and a half it has operated in a truly hideous environment. Not a proper server room. Not even a server closet. A crappy rack in a neglected corner of a customer premises in an industrial area with unbelievably marginal power.
Despite regular brownouts, blackouts, misconfigurations, customer stupidity, my stupidity, cabling issues, dead switches and more the Tintri T-850 has proven to be the single most reliable centralized storage device I have ever used in my entire career. It is rock solid. It is beyond fast. It is a damned fine machine. I cannot thank Tintri enough for the loan of the device.
In January, I get to recover the Tintri from the customer site. In January I am also building a public-facing cloud offering to meet the needs of my customers. With any luck, I’ll be able to buy the unit off of them and make it the core storage of my cloud. I can think of nothing I’m more comfortable betting my business on.
Thank you, Tintri. Your storage Just Works, and that has already saved me endless grief.
Micron get on my thank you list for two reasons. The first, they gave me a giant pile of NVMe SSDs for testing and this has hugely transformed my lab, the things we’re able to test, and my understanding of how queue depths affect storage. The ability to play with these drives (which I could not otherwise afford) has made me a more competent storage administrator, and that alone would be a huge thank you.
While “free stuff” is nice, the thing that I really want to thank Micron for is making SSDs that don’t randomly crap out. It’s been a bad year for me with storage. I’ve had a huge number of both mechanical hard drives and SSDs go. These range from the SSDs in my workstations to an entire bank of spinning rust in my NAS.
The drives that haven’t died are made by Micron. Some are Micron branded, some are Crucial branded, but the it’s my Micron-manufactured units that are the ones that just keep going.
I realise that it might be sample size bias at work to ascribe better outcomes to the Micron devices. I don’t operate at hyperscale and I’m not going through tens of thousands of devices a year. But this year, in my little corner of the world, I have gone through several hundred devices of a few thousand in service, and it’s the Micron ones that have yet to die.
Thank you, Micron. It’s nice to have something to put in my workstations and servers that my data stays safe on.
If I’m busy thanking people for free stuff than Supermicro has to be on the list. A lot of Supermicro gear goes through my labs. Some of it isn’t worth the hassle of shipping back, so we get to buy it off Supermicro for a song. A decade ago, this would have made me deliriously happy. Today, I have an entire lab full of gadgets and gizmos, so I am in many ways more worried about where I might put new things than I am deliriously happy about getting them.
That said, Supermicro deserves to be on my thank you list for two reasons completely unrelated to their assistance building up my lab. The first is simply the continual stream of units that come through my lab, especially the stuff so awesome that it’s more than worth shipping back.
Supermicro is at the forefront of new hardware tech. They had Xeon-D-based servers out before their competitors, units with exceptionally long design lives that proved to be wildly popular and have changed both my lab and the production environments of my customers. They have switches and microblades, “Twin” servers and gaming motherboards.
Supermicro’s constant dance on the cutting edge has allowed me to keep up to date with the latest technology. As someone who has to write about it, this constant exposure is fantastically useful, and something I couldn’t get any other way.
The other reason I have to thank Supermicro is the quality of the gear they make. Not only has it proven to be exceptionally reliable, they don’t charge extra for their IPMI or IPKVM implementation, and I can’t thank them enough for that.
The Supermicro side of my lab is something that I can turn off or on as needed (important when you don’t have enough power or cooling in the server room to power it all on at the same time!). It’s also something that I can log into remotely to load up operating systems, change configurations and so forth.
Enterprise sysadmins might take that for granted, but in the small business (or home lab) market, full IPKVM functionality is a godsend that we don’t always have the money to pay extra for.
Thank you, Supermicro, for making long-lasting equipment that does what I need it to do out of the box.
IoFABRIC are a startup that actually listens to others. They evolve their product based on feedback they solicit, both from industry wonks and from customers.
Unlike so many other startups opt there, ioFABRIC aren’t charging around thinking they know better than everyone else. Their vision, while important to the development of their product, isn’t the be-all and end-all. They accept that they might not get priorities right, that they might miss a feature or that they might have create a bug.
So much of my grief with vendors comes not from having to fight the product, but from having to fight the people. Overcoming the arrogance of individuals and corporate hubris as a culture in order to get them to even be willing to try to understand my needs or problems is usually a huge thing. Not so with ioFABRIC.
Thank you, ioFABRIC, for caring about more than just the imagined and idealized IT needs contained in your own memories.
Devolutions make Remote Desktop Manager (RDM). For a systems administrator, RDM is $deity-tier.
RDM allows you to, from a single application, connect to virtually anything. RDP, SSH, VNC, Azure, PowerShell, Citrix and on and on. And on and on and on and on and on.
With RDM you can save your credentials, create groups of things to be accessed, share your configurations with co-workers, automatically VPN or SSH into remote networks as part of the connection, and more. Think about the ability to right-click on a folder, go “open group”, and launch 50 tabs worth of connections into various servers that you need to talk to every month for maintenance.
I know that at scale you’re supposed to use Puppet to do everything. Desired state config and infrastructure as code, etc. But I live in a world of SMBs. Lots and lots of SMBs, and it’s super extra neat-o to be able to launch connections into all my customers’ servers at the same time to do the monthly patching.
Thank you, Devolutions, you make the daily grind a lot less crap.
It would be improper of me to write a thank you blog without mentioning Ninite. Ninite is my friend. It is my digital comfort blanket. It is that tool which I fall back on so frequently that I feel almost naked without it.
Ninite comes in two flavours: the website and the application. The website allows home users to select a bunch of check boxes and get a generated installer that installs the latest versions of whichever popular third-party applications they chose onto their computers.
The Ninite Pro application is a paid application. Like the website, you can select from popular third-party applications for install. You can also use the application to upgrade any of those applications that might be installed.
More critically, Ninite Pro can talk to Active Directory, find every single Windows system on the network and then update the third-party applications on all of them. A couple of button pushes and Java, Flash, Adobe Reader and other horrible security nightmares are up to date across the board.
I use it to uninstall things like Adobe Reader and install Sumatra instead, or to install Chrome and Firefox without even having to use IE. (Now IE really has no purpose!)
Ninite doesn’t have everything. The number of third-party applications it supports is actually rather limited. That said, it has the most important applications and is exactly the right tool to make keeping the systems of friends, family and small businesses up to date.
Thank you, Ninite. You make the holiday grind of sitting in a corner during family gatherings fixing everyone’s notebooks a lot less miserable.
To everyone else, vendor or not, who makes life that little bit better for others: thank you…and Happy Holidays.
- Information Overload? There’s an app for that. - January 12, 2017
- Year end thank yous - December 23, 2016
- Archival cloud storage can be an affordable backup layer - October 3, 2016
- On the importance of the user experience - August 13, 2016
- Beyond the traditional storage gateway - June 17, 2016
- Data residency made easy - June 15, 2016
- DevOps shouldn’t be a straitjacket - March 15, 2016
- Preparing for Office 2016 - November 7, 2015
- Supermicro, VSAN and EVO:Rail - February 4, 2015
- Make a #WebScaleWish - November 21, 2014