From where I sit, things have gotten a lot better in information technology recently.
Open source has a lot to do with that. Users try the software - and expect to see the software and be able to fix and enhance and, yes, fork the software themselves.
Users can decide to move along when and how they want to move along. And that gives them confidence in using a project in the first place. We certainly believe in this at MayaData - and we also believe in open governance, which is why we’ve now contributed two thriving projects to the Cloud Native Computing Foundation, OpenEBS, and Litmus.
Here’s the thing, though. What about when the user needs help to operate Kubernetes as a data layer, with the help of OpenEBS or LitmusChaos or both? Maybe they want 24x7 support, per workload back-ups, off cluster logging, and an enterprise edition for long term support. Then they come to our Kubera and try it out for free or on a low cost per user license. Pretty quickly, they wonder - how much will this cost when Kubernetes really is our data layer. Instead of one user and, for example, 5 clusters and 10 worker nodes storing data, what happens when we have dozens of users and thousands of clusters?
We have not done a great job to date of explaining our pricing. My goal is that our pricing should be 100% open, just like our code - thereby ensuring we remain aligned with our users and customers. Pricing should tie us to the success of our customers. And, too, pricing should reflect the value of the software and the services we deliver.
Pricing - fundamentals
Get started - for free:
Kubera starts with a free tier, and of course, the components of Kubera are themselves open source. That should be pretty straightforward. Free = free.
Individual and team licenses:
The pricing for Kubera includes per-user licenses. If you help run workloads on Kubernetes, we think you’ll find Kubera to be helpful. For $79 per month, you can grab all of Kubera's functionality, including 24x7 support for OpenEBS and eventually for Litmus as well. We are running a discount for a limited time for $49 per month if you sign-up now.
Per cluster or per node licenses:
As your usage increases, you’ll graduate into a scale of operations that users have asked us to license on a per worker node basis. If a worker node is used by OpenEBS to store data or manage stored data, then that is a licensed node.
Here we simply price at $2,500 per storage node per year.
True-up processes:
Okay, but what if you don’t know how many nodes of Kubernetes and storage you will run in a given year? After all, on the cloud, we simply pay for what we use - or at least for what we provision whether we use it or not.
Our approach is simple - we track usage or ask you to tell us your usage if you are running in an air-gapped environment - and each quarter, if your usage has exceeded the amount you have licensed, then we send you an invoice. And here’s the kicker - you have a 30 day fix period.
We started off trying to emulate cloud pricing. Many users tell us we’ve done even better by providing this fix period. That way if a customer spins up a huge number of stateful workloads towards the end of a quarter, you are not punished if that was a short term peak.
This helps keep us all aligned and, importantly, helps when it comes to budgeting.
Multipliers and discounts:
Enterprise pricing:
For some use cases, you may want tighter SLAs, proactive consulting, and more. If you want our highest level of possible support, and you want us to run a version of your environment at MayaData so all releases are tested against it and so we can help you do what-if analysis and capacity planning and more, then your cost is 1.75x that of our base pricing.
Prepayment discounts:
On the other hand, you can also drop your pricing considerably by committing upfront to an annual license and paying upfront. Keep in mind that your actual usage will be “trued up” via the true-up process explained above - so your risk is less than you might think.
The prepayment annual license discount is 20% per year. What that means is prepayments for three years = 60% off. Now, this assumes that your purchasing department actually, you know, pays.
A brief digression on purchasing departments - they are Larry Elison’s fault. Ok, that might be an exaggeration. After all, if you know me, then you know that I firmly believe that “Larry is a brilliant and proven technology executive.” The point is that purchasing departments exist in part to protect technology users from their vendors because their vendors have exploited them far too often in the past.
We’re not exploiting our customers.
If your purchasing department has to delay payment - because clearly a Global 2000 needs help from a small company like us to finance them - ok, great. We charge 5% per month beyond the first 30 days. Sorry, not sorry.
Site licenses:
Some of our largest users use us on a huge scale and don’t want to worry about how much they will pay ever again or at least for the next few years.
So we came up with site licenses. Many of these users are at a scale of 4,000 or 5,000 nodes with some component of OpenEBS running on them, and some are deploying at an even larger scale. With that in mind, we came up with a price level of $1.2 million per year for a prepaid site license. For a particular customer, this represents in excess of a 90% savings versus per node pricing.
Conclusion and Return on Investment:
This short blog is intended to help clarify our pricing. In our open source approach and our open business model, we are trying to eliminate the risk of a gotcha moment - the very sort of lock-in based gotcha that, unfortunately, cloud vendors and especially storage vendors are using to grow their businesses often while causing angst on the part of their customers.
Which gets to ROI or return on investment. One value of extending Kubernetes to become your data plane with the help of our container attached storage and Kubera is you are now back in control and, as such, less at risk of predatory pricing and gotcha bills from your cloud and systems vendors. Yes, you’ll have to negotiate with them to get more fair pricing and share some of the ongoing drops in their hardware and bandwidth costs. Ask for 50% per year is my advice - that’s approximately the rate of improvement of their underlying hardware's price performance. With the help of MayaData and Kubera, you’ll be in a much better position to negotiate.
Other sources of ROI are pretty clear. You can simply move faster, run workloads in a more loosely coupled way, and hence operate your software in a much more agile way. You can also use locally attached disks in some cases instead of expensive and poor performance shared everything storage or cloud volumes - or just use those underlying expensive systems more efficiently with our help.
Our pricing will probably change over time as we learn more and our products and projects continue to grow to meet the needs of users. Nonetheless, we commit to remaining transparent so we can truly partner together to help customers achieve incredible outcomes by using Kubernetes itself as their data plane. We also want to make costs more clear, more transparent, and more tied into Kubera so that you can see, for example, your ROI from using Kubera in more real-time. Any and all feedback is welcome and definitely needed.