Some thoughts on CDNs
Written on 5 Sep 2016
The greatest touted advantage of a CDN is faster load times. Because CDNs have worldwide servers, users can fetch a copy of your content from a server nearby which should be faster.
But is this always true?
What if my website specifically targets Dutch users and my server is in the Netherlands? Is it still faster for all my Dutch users?
CDN performance may not be consistent. One particular location may be blazing fast, and another may be very slow (and how do you know which locations are slow?)
A CDN also introduces performance overhead in the form of a DNS request, a new TCP connection, and possibly new TLS negotiation; so a CDN has to not only be faster, it has to be fast enough to offset for this.
Average load times are nice, but what about the worst possible load time? In my experience, this is often a lot worse with CDNs. Like with most things in life, a single ‘average’ number is pretty useless.
CDN performance is pretty hard to measure. In fact, I can’t really find any good figures on the web that aren’t produced by a CDN provider (and thus not reliable). My personal experience is checkered, and while a CDN can most certainly improve the load performance, I would be careful in just assuming so.
It is already very easy to serve static content with very high performance using tools such as Varnish. In many cases, this is about as fast (and certainly ‘fast enough’).
Any server can be compromised, no matter how secure. It is entirely possible for someone to breach your CDN and inject malicious code.
Stuff like this does happen, for example with Linux Mint this week and externally loaded malware ads are also a thing. There are many, many, more examples to be found.
These events are comparatively rare and usually (but not always!) quickly discovered. But they do happen. It is a risk, and using a CDN increases the attack surface.
An important question here is what sort of website are you developing? What is the maximum damage that can be done? For example, the Plesk interface loads a Facebook ‘like’ button by default, effectively giving Facebook complete control of your server if they decide to take it. The chances that they actually will do this are small, but the potential damage can be huge, and IMHO it’s not worth the risk.
But if you just serve a basic weblog or some such, the worst that can happen is that you serve ads or malware for a while (probably along with thousands of other sites). Not great, but not a disaster either.
In a similar vein, the CDN provider may “go bad” in the future and start inserting adware or such in the CDN. Sourceforge does this with some downloads.
This is chiefly a concern with public CDNs and quite rare, but it does happen. Are you sure that bootstrapcdn.com won’t start serving crap in ten years’ time?
This is especially a concern for sites you write once for clients and then never look at again. Ideally, they should keep working for as long as possible.
The old adage of ‘everything is a fucking DNS problem’ has been replaced with ‘everything is a fucking caching problem’.
Caching is great, but it also adds a lot of complexity. CDNs effectively add another layer of caching, and things can – and do – go wrong here.
Think about the drawbacks, try to optimize your server with Varnish first, and please, measure the actual real-world performance (which is probably easier said than done).