2025: The cycle never ends
Another year has come to an end. It has been another wild ride. Here is the yearly recap. Also check out last year's recap if you're interested. That's incidentally also the previous post on this blog.. I have not been very motivated to write anything lately.
Statistics
What most people are probably here for.
The number of patrons increased from 13042 to 17446 (34% increase)
Monthly earnings through Patreon increased from €41672 to €64611 (55% increase)
1 259 534 TB of bandwidth was used (105% increase, that's 1.2 Exabytes!)
2 682 977 991 files were viewed through the website (23% increase)
2 339 944 253 files were downloaded (1% increase)
Bandwidth capacity was upgraded from 900 Gbps to 1600 Gbps (77% increase. It takes 5 seconds to download a terabyte of data at this rate)
Storage capacity was upgraded from 11232 TB to 25056 TB (123% increase).
I have committed 132 changes to the Pixeldrain server repository, written 6566 lines of code, and deleted 3768 lines of code
Even more storage servers
Shortly after migrating all the files from Hetzner to GSL last year the storage space was running out again. I ordered another full rack of storage nodes. We learned some things from the last deployment, and one of those things is that Pixeldrain runs better on a large number of relatively smaller storage nodes, as opposed to a few really large ones. It's hard to scale a storage system into the thousands of hard drives range, and I'm starting to notice a lot of issues pop up. This year the storage server performance has been the primary limiting factor in Pixeldrain's growth, and I have not been able to find a good solution yet.
The new storage cluster consists of 21 servers with 24 disks each, 12 in the front and 12 in the back. The server chassis are SuperStorage 6029P-E1CR24L which have two layers of drive mounts, providing huge storage density. The first server rack also had the issue that it was too dense to fit any routers in there, so in this one we left some gaps.
This cluster has been running for close to a year now, and we are once again approaching full capacity. This time I can't invest a huge sum of money into a new cluster I'm afraid. I'm still financially recovering from the last expansion. Instead I will have to start reducing the storage limits for free and Pro users. The days of unlimited storage for €4 per month have to come to an end soon.
Stagnated growth?
The view and download counts seem like they have not really grown this year. This is probably due to some restrictions I have added to avoid platform abuse. Abuse is something that has only gotten worse over time. People are constantly trying to find ways to circumvent Pixeldrain's limits. Although it's not something I want to spend much time on, as it's not really my idea of fun, I was forced to introduce some new restrictions this year in order to keep my caching servers from exploding.
The largest change was the removal of iframe embeds for free users. This was being used to register fake views on files which prevented the hotlink detection from working.
I also blocked the usage of download limit bypass sites using Cloudflare workers. Cloudflare provides free unlimited bandwidth to their worker scripts. And you can choose to run each worker on a new IP address. This allowed people to get a fresh download limit for every file they downloaded. This was pretty easy to block as Cloudflare adds HTTP headers to all outgoing requests. You can still use Cloudflare workers from premium accounts.
Deplatformed once again
Around march my payment provider Mollie decided to spontaneously disable my account. I have asked them what the reason behind this decision was a few dozen times (they're located in the Netherlands, so it was easy to call them). But they simply wouldn't budge. They say my account had to be terminated due to internal policies, and that's all they have to say about it.
Around the same time a representative from PayPal approached me to ask if I wanted to integrate with their payment system. Since I immediately had a direct line of contact to a real person working at PayPal it seemed to me that this is a safer option than hunting for new payment providers again. PayPal also offers payment gateways for regional payment processors, including credit/debit card. But their API is not that easy to work with and I have not gotten it to work yet.. Eventually it will be done.
Legal shenanigans
This had to happen eventually. The British telecom regulator Ofcom contacted me with regards to my compliance with UK telecom laws. I went on a video call with them and they told me they were preparing a list of changes I would have to make to the platform. While they seem pretty friendly in person, every e-mail I receive from them has a note at the bottom stating the ways they would punish me if I don't comply with their demands. Needless to say I have little choice in this matter.
The first change I need to make is content scanning. I need to implement a system to visually compare images and videos on Pixeldrain to a CSAM database. I know how to make this, but it's still quite technically challenging and I will need to spend quite a lot of time developing this. I will also have to order dedicated servers for this scanning job as it's not particularly easy on the CPU.
Secondly I have to change my TOS to contain more strict definitions of prohibited content. I am not really well versed in legal texts, and hiring a lawyer to set it up for me is not something I can afford. I might have to resort to ChatGPT to solve this for me. Same goes for the third demand. They want me to write up a risk assessment report. I am honestly not even sure what that means. I'll have to research it.
How I feel about all this
Once again my hobby project is reaching an unprecedented scale. Still I can't help but feel a little bit disillusioned. All this growth once again brings a lot of stress with it. I constantly feel the obligation to keep improving things for the sake of the tens of thousands of paying people who use this site. Yet it's getting harder and harder to keep finding solutions to the issues I'm running into. Things are always breaking (luckily not during my vacation this time). More and more often it's due to factors outside of my control. Whether it's third party APIs I depend on shitting the bed, or governments censoring more and more of the once-free internet. 2025 was really the year of government censorship, and it does not look like 2026 will be any better. I guess fascism is back in style.
I am not sure how much further this platform can grow when I'm the only one maintaining it. It might be time to shrink the platform slightly. The only way to do that is by introducing more restrictions on the free tier, which will inevitably make people angry.
In order to distract myself from working all the time I have started sinking more time into my hobbies. I got more serious about running, and my physical health has been improving as a result. I also got much better at playing guitar.
A new project
I alluded earlier in this post that I may have to shrink Pixeldrain to keep it sustainable. The truth is that I have already been working on a plan for the future. My idea is to take all the code and knowledge I have built up from 10 years of working on Pixeldrain and starting a new, more professional, cloud storage product.
The new project will pretty much be a standalone version of Pixeldrain's premium filesystem. It will be paid-only and there will be no free tier. I understand that this will turn off a lot of people, but there are multiple reasons for going this route.
The payment requirement increases the stakes for people wanting to abuse the service. I will be quite strict in enforcing my TOS, and if you get banned you also lose your account funds.
Premium-only also means that there is no server load caused by free users driving the prices up. On Pixeldrain 90% of the traffic is used by free accounts. This represents tens of thousands of monthly euros in caching and storage capacity. All the money to support this hardware comes from you, my patrons. With the new service there will be no free tier to support, and in turn the product can be priced a fair bit lower.
The new project will most likely be called FNX.storage. I have been trying to think of a better name but nothing comes to mind. And I have already paid for the domain (it's three letters.. expensive!). There are still many changes that need to be made, both in the front and in the back, to make this project truly scalable. I have learned from Pixeldrain's issues and would like to get it all right this time.
FNX.storage will most likely launch in the first half of 2026. When this happens, you will hear from me on this blog.
Happy new year!
When I started writing this post I thought it would be a short one, guess not. I may have had more to say than I thought. Enjoy the fireworks tonight. And know that I love you all.
Greetings,
Wim.