Server options - pros and cons

Sitebulb Server is designed to be an 'always on' solution and we have chosen our server partners with this in mind:

  • Ionos - offer fantastic value with great spec machines
  • Prostack - offer top of the range machines with first grade, fully-managed support

But there are of course other options, and some of these might be more accessible or familiar to you, for example:

  • You can set Sitebulb up on a spare Windows desktop machine on your network
  • You can spin up an AWS instance and run Sitebulb Server
  • You can spin up an instance of Google Compute and run Sitebulb Server

In this article I will run through the pros and cons of the different options so you can make an informed decision.

Dedicated Servers

For example, renting a server from Ionos or Prostack on a monthly or yearly basis (note, we have a video guide for Ionos setup).


  • A completely dedicated machine just for Sitebulb Server - no resource sharing or virtualization happening.
  • Compared to cloud solutions, dedicated servers are significantly cheaper - assuming you wish to run the server 24/7.
  • Costs are fixed and capped, meaning you know ahead of time how much you will need to budget for your server.
  • Each core you pay for gives Sitebulb access to 2 threads (so an 8 core machine gives you 16 threads).


  • You can't spin up and spin down, so if you only wanted Sitebulb Server for 1 week (e.g. a one-off enormous crawl), you'd end up paying for server time you don't need.
  • Can take hours or days to get set up.
  • Since it is a physical box, you could suffer a hardware failure that would cause downtime to resolve.

Cloud Instances

For example, spinning up an instance in AWS or Google Compute (note, we have a video guide for AWS and a video guide for Google Compute setup).


  • Quick and easy to spin up, can get going in literally minutes. Similarly, it is easy to switch to a more powerful solution if required.
  • You can spin up and spin down, so if you only wanted Sitebulb Server for 1 week (e.g. a one-off enormous crawl), this would be an ideal fit - just terminate the instance after you are done.


  • If you keep the box on 24/7, they work out very expensive - much more expensive for comparable dedicated servers.
  • Cost can scale up based on usage, making budgeting more unpredictable. 
  • Since they use VCPUs, you only get 1 thread per core (so an 8 core machine only gives you 8 threads)

Spare desktop machine

This is only an option if you actually have a suitable machine at your disposal, and it can be made accessible as a server (i.e. you are able to set up port forwarding through your network firewall). To many people, this simply will not be a viable option.


  • You don't have to spend any extra money on renting a server.
  • You can get going straightaway.
  • Makes use of an otherwise redundant machine.


  • In most cases you will just be using whatever machine you've got, so you may not be able to get the most out of Sitebulb Server.
  • Crawling on your local network will use up bandwidth.
  • If you need to switch to a more powerful solution, this will mean buying, installing and managing the hardware.
  • If the hardware fails, you will be responsible for fixing or replacing it.

Sitebulb's recommendation

In general, this is a case of 'horses for courses'. If you have access to a spare machine, and just want to give Sitebulb Server a try, then this is the perfect way to test it out. You may also then find that the 'spare machine solution' actually works for you long term.

If you don't have a spare machine, but do want to test out Sitebulb Server, then spinning up a cloud instance is a great way to try it out. You can get an instance up and running (with Sitebulb Server fully connected) in only 10 minutes - see our video guide for AWS or our video guide for Google Compute for instructions on this. You can keep your instance on for a week or two and kick the tyres a fair bit, then turn it off when you are done testing.

If you were planning to use Sitebulb Server on an ongoing basis, however, then AWS/Google Compute is a really expensive solution - this is where dedicated servers really shine, and we recommend Prostack as a premium, fully managed option, and Ionos as a budget-friendly alternative. 

Further reading