Realize cloud savings without adding tools and FinOps
Cloud cost management has become popular, and for good reason. Companies are often able to cut their cloud costs by upwards of 40% - but realizing the potential on traditional clouds is very complex.
On Merrymake it's simple - you only pay when your code runs, not for over-provisioning and reserving resources.
(and if you already know, speedrun here).
The large savings potential comes from how traditional clouds are provisioned and their billing structure. On a traditional cloud, companies rent a fixed amount of server capacity, specified in what is called an instance. To ensure that the required performance is always available, they often end up with over-specified and under-utilized cloud instances.
As an example, imagine a large batch process, like a payroll process, which runs once a month. On a traditional cloud, companies provision an instance with the amount of resources that'll give the required performance. But to make sure, that the payroll runs when it's time, there's only one thing to do: reserve the instance and its resources by paying for it, also when nothing is running.
This goes for all applications that experience spikes in activity or traffic. IoT is another example, where activity can go from zero to a lot and back to zero several times a day. Again companies have to pay for being able to handle "a lot" even during periods with little activity. Because it can be hard to know what "a lot" is, companies often over-purchase rather than taking the risk, that a sudden spike could exceed monthly quotas and break applications. On top of that, for user-facing applications, companies also have to pay for keeping services warm, so users don't experience poor performance after periods of inactivity, because traditional clouds struggle with cold starts. See how costs are building, and how many of them are not even spent on actually running code?
This is the discipline of trying to optimize cloud provisioning on the performance-vs-cost-continuum. On traditional clouds, companies need to do two things:
Does it sound complex, both technically and organizationally? That’s because it is. Sounds like it'll spend developers time on not developing software? That’s because it will. And then imagine doing this in multi-cloud environments..! Nevertheless, it does work for a lot of companies - but the same can be achieved much simpler. As teased above:
We’ll just leave that standing there.
... and then elaborate on what this means in real life. On Merrymake there’s no need for provisioning instances, no stepped upfront cost as your services scale, and no wasted costs for reserving resources or keeping services warm, as you would have on a traditional cloud.
Merrymake simply runs your code when it's required and stops when it’s not. And when your demand scales up and down, Merrymake does too, automatically and instantly.
This is possible because Merrymake is truly serverless. This means, among other cool things, that we have zero cold starts, so you can maximize performance and minimize cost at the same time.
In other words, cloud cost management and its inevitable complexity is a thing of the past, when running on Merrymake. Returning to the payroll example from above, a switch to Merrymake would automatically realize the below cloud cost reductions:
Exemplary illustration
And as the icing on the cake, Merrymake also stands out from other platforms by offering free network traffic. Many other platforms also price their services based on network traffic, which can be unpredictable and challenging to calculate accurately, leading to bad surprises when the bill arrives. Not on Merrymake.
