Last year, after about 7 years in the Database Administration world, I decided to transition to a Software Engineer role. I needed a project to dive into to help get up to speed and ready for that transition. I decided to tackle a pain point in one of my hobbies.
I’ve been involved with paintball on and off since I was in my teens. One thing has always been consistent – a strong used equipment market where players buy, sell, and trade gear. Back when I first got into the sport most of that activity took place on forums. Over the years that community has mostly moved over to Facebook groups. Both have some strengths, but I felt both were lacking pretty significantly.
Honestly, I still prefer the forums to Facebook. The most
popular forum has a feedback system that, for the most part, helped users
confidently do business with each other. I view that as the most valuable
feature of that platform and it is what always keeps me coming back.
Unfortunately the forums haven’t changed much since the early 2000’s. Images require third party hosting and BBCode to share. Chat functionality is pretty primitive. Facebook wins by a landslide in the usability category and that has resulted in a shrinking user base on the forums.
Facebook’s biggest flaw is that it actually doesn’t even allow for paintball markers to be sold on their platform as it categorizes them as firearms (eyeroll). Because of that, users don’t actually use the marketplace features of Facebook. Instead they meet in groups and arrange deals privately over Messenger. Groups and posts get flagged/removed all the time and it’s hard to keep up with the current relevant groups and their rules. This also circumvents the feedback system which makes scamming even easier.
Ebay isn’t much of a contender here. Their fee structure scares most users away and a lot of the gear you find is overpriced to compensate. There’s also no real way to trade. I’ll use it from time to time, but it’s always my last option.
I set out to build a platform that combines the strengths of all the above while keeping fees minimal. A modern platform with a simple UI that drives a transaction workflow capable of facilitating all types of deals. By building workflows around the various transaction types, it removes all doubt about what is expected of a user and eliminates the need to figure it out deal by deal.
High level feature overview:
Offer/counter offer system
Ability to handle standard purchases, purchases from cash offers, straight trades, trades adding cash, trades including multiple items.
Modern chat functionality
Ability for sellers to offer free, fixed, or estimated shipping (USPS/FedEx/UPS)
Admin panel to simplify moderation
Closely integrated with PayPal to mirror current user experience
Hosting/maintaining this wouldn’t be free and I’d need to monetize this somehow. I decided a minimal flat fee is how I would tackle that. While it’s possible that users could reject the idea of an additional fee, I believe the value add is there.
I already mentioned that Snakefifty uses PayPal, but I did consider other options. I would have loved to use Stripe Connect (their marketplace offering) but the pricing was just unrealistic for this niche. Fees would have ended up being similar to eBay and that would be a non starter. I ultimately stuck with PayPal as that is what the vast majority of the community uses and is already familiar with. I simply leveraged PayPal’s parallel payments to collect a fee.
I am pretty proud of the platform and I believe it has the potential to provide significant benefit for the paintball community. That said, I’ve decided to stop work on it for now to focus on other projects/interests.
Taking a quick pulse of the community it seems that while many acknowledge the pain points in the current channels, most seem tolerant of them. Acquiring enough users to make the platform viable is a significant challenge and one that I am not entirely confident I would succeed in. Even though I aim to keep fees minimal, I can’t get around the fact that I am adding a fee that doesn’t exist on Facebook and the forums.
There have also been some legal/tax changes that complicate running a platform like this. In 2018 the Supreme Court ruled that states could charge sales tax on out of state purchases even if the seller doesn’t have a physical presence in the state. That in itself wasn’t news to me. What I didn’t realize, however, is that over the last year or so most states have shifted the burden of collecting/remitting taxes to “marketplace facilitators.” The benefit is that an individual seller, for the most part, no longer has to worry about paying taxes to every single state their customers reside in. States also save on resources as they don’t have to try to collect from individual sellers.
Putting that burden on the marketplace seems reasonable in the cases of Amazon and eBay, but it makes it significantly more difficult to run a small marketplace like this. While most states do carve out exceptions for marketplaces with limited volume, if I ever cross those thresholds it will turn into a full time job to keep up. Keeping fees minimal is a hard requirement for this platform to be viable and even with the volume to cross those thresholds it wouldn’t be worth the effort required. Even if I could find a way to keep up, users would likely perceive the tax as just another fee that they don’t have to deal with when they deal privately in other channels.
Ultimately I went into this project knowing that it would be a pretty big challenge for it to succeed as a business. That was a secondary goal to building something that let me get up to speed in my new role. This project was absolutely a success on that front.
Here’s a quick tour of the project in it’s current state:
I made some fairly significant upgrades to my home lab a few months back and an update is overdue. Without wasting any time, here is the 2019 edition in all its glory:
My previous setup did what I needed it to but the storage situation was a mess. Lots of standalone drives with no real redundancy. I initially looked into a few options for addressing this including buying a NAS and building a NAS. Then I stumbled on Unraid. I was vaguely familar with Unraid but I never gave it a real look until recently. Unraid is super flexible with storage options, has a built in hypervisor, and makes standing up containers super easy. Ultimately the relatively low cost of used hardware and the flexibility of Unraid was the best combo for my needs. My plan was to build out a new server, shuck all the external drives, and migrate everything off of my ESX host (all VMs) to the new Unraid box.
After stalking eBay for awhile, I ended up with the following:
2x Xeon 2650 v2 (16 physical cores)
64 GB ECC DDR3 RAM
GeForce GTX 1060 6GB SSC
46 TB spinning, 1 TB SSD (already owned the storage)
Fractal Design Define R5 case
The migration itself wasn’t trivial and it made me question why I created so much work for myself outside of work. I managed to get everything completely migrated to new containers/VMs on the Unraid server, with the majority of that going to containers. This also included migrating the physical Windows server that was running on the TS140. That newly freed up TS140 was repurposed into my development server. I am currently running Proxmox and a handful of VMs on there. I had a UPS die in the middle of all this (found out during a power outage, as seems typical) but luckily with the reduced power draw compared to the old i7 I can handle running everything off of one UPS.
You may have noticed I threw a pretty nice graphics card into a server. One of my favorite things to come out of this is the capability to do remote gaming. I have the GPU passed through to a VM on my Unraid server and I use NVIDIA GameStream and Moonlight to game on my Thinkpad. It has been pretty much flawless to this point. I am not a huge gamer and the GPU helps with some other interests (editing and such) but it has given me the ability to play modern games on high settings without a dedicated gaming computer.
No changes on the networking side of things other than that I now run all my home DNS through Pi-hole.
I’m a big believer in learning by doing. Over the last couple years I’ve built up a nice little home network/lab where I host various applications and test new tech. I will hardly be the first person to make this recommendation, but if you are able to get your hands on some hardware to set up a home lab I would highly recommend it. Any old computer you can turn into an ESXi host is enough to get started. I have learned an enormous amount about various technologies that I wouldn’t have had much opportunity to learn elsewhere. First, the obligatory picture:
Ubiquiti Edgerouter X
Ubiquiti UniFi AP AC LR (not pictured as it’s located in a location central to the house)
ARRIS SURFboard DOCSIS 3.0 Cable Modem (SB6183)
TRENDnet 24-Port Cat6 Patch Panel (22 runs in the house, minimum for 2 per room)
Startech 12U 19-Inch Desktop Open Frame 2 Post Rack
Raspberry Pi 3 Model B
Xeon E3-1226 v3 3.3Ghz
16 GB Ram
44 TB (raw) of various spinning storage, mostly WD reds (yeah, yeah, I know. I’ll get a NAS eventually)
Old computer serving as my ESXi host
Intel Core i7-2600K Sandy Bridge Quad-Core 3.4GHz
16 GB RAM
Samsung 840 EVO 250GB SSD
5 TB (raw) of WD Reds
2x CyberPower CP1500AVRLCD UPS
One is for my “core” equipment (network stuff/TS140), the other is on the ESXi host
I am a huge fan of the Ubiquiti hardware. This combo was one of the first purchases and its pretty fantastic to not have to regularly “reset the router” like seems to be necessary for just about every consumer router/AP. I picked up both switches off of eBay and have been pleased with them. Some Cisco purists dislike the small business line but they’ve been great for my purposes. The SG300-28 (gigabit) is my main switch and the SF302-08MPP (fast ethernet, POE) runs my cameras. I didn’t need to have separate switches, but based on what was available on eBay this was perfect for my goals. If we’re being honest, need went out the window a long time ago anyway. I run a handful of VLANs and 2 separate physical LANs to segment off the different portions of my network. Necessary? Probably not, but where’s the fun in that? This has allowed me to tinker with VLANing and setting up firewall rules. Fun stuff!
The TS140 runs AD, DNS, Blue Iris (cameras), and acts as a file server. My storage situation could certainly be improved, but all important data is backed up locally and remotely and any data that I consider to be less important is at least mirrored to separate local storage. My Raspberry Pi runs PiVPN which is a super simple way to run OpenVPN and is much more manageable than my previous setup where I ran it on my router. The ESXi host runs the following VMs and applications:
Plex and related apps
Kali box for security tinkering
Windows VM for when I need to run Windows specific apps
Whatever I feel like playing with at the moment
Each of my Linux boxes are running Ubuntu 16.04. All VMs are regularly backed up using ghettoVCB.
That’s pretty much it for now. I will update when I inevitably add more!
In January I posted about a couple of upcoming speaking engagements and after some nagging for a follow up from @JayDue216, here it is. First up was the Ohio North SQL Server User Group’s January meeting. This wasn’t my first time speaking in a professional setting but it was by far the audience with the most SQL Server experience I have been in front of. All in all, I thought it went alright. I didn’t feel too great about it immediately afterward but I received some good feedback from those in attendance which is always great to hear. Nerves got to me a bit and I caught myself tripping over my own words at a few points. I had a great deal of confidence in the content I developed and I can’t express enough how valuable that is when your delivery isn’t going as smoothly as planned. Having that confidence allows you to slow yourself down, gather yourself, move to the next slide, and keep powering through the session. At the end of the day it felt good to get the first one out of the way and identify areas for improvement.
After getting my feet wet next up was SQL Saturday #357 in Cleveland. I was initially scheduled to speak early in the day but an out of town speaker was hoping to catch an early flight home and I agreed to move to the last time slot of the day. I caught a few sessions early in the day, but I ended up spending a great deal of the time in the speaker room going through my presentation to make sure I was ready to go.
I only did it for the sweet SQL Saturday pullover
Unlike the user group, I felt very good about the presentation after I had finished. I didn’t feel myself tripping over my own words this time and my demos went smoothly. Despite being the last session of the day I had a full classroom and the audience seemed pretty engaged throughout the entire presentation.