[Hosting Services] Most Important Question To Ask

[MJ]

I don't think choosing web server is the end of the world. The only advantage of Litespeed I think is, it is optimised out of the box. The comparisons over the web are with default installation of Apache/ nginx with LS. If you fine tune, nginx, it will perform quite similar to LS.

Just have a look at this comparison - https://centminmod.com/benchmarks_nginx_openlitespeed_cherokee.html
https://wpjohnny.com/nginx-vs-openlitespeed-speed-comparison/Centminmod is fully optimised lemp stack.

There is slight difference in performance between the two, this is what happens if you fine tune nginx. You can even use Apache + nginx, nginx used for serving static resources and apache for htaccess support.

The advantage of using Litespeed is, you get free caching plugin from LS which includes tons of features. For other servers, you will have to rely on paid plugins.

 

[Dopious]

Okay, and as for everyday individuals with an average Wordpress website that would never benefit from this metric

In my humble mind, Litespeed is great, BUT the ones that benefit most from it is the hosting owners, since they can load up their servers with more customers/accounts. The regular brick and mortar company will not see those huge gains for a website that "might" get a couple of hundreds of visitors per day. I use both Litespeed and NGINX servers and CPU SPEED/MEMORY are more important for my customers. Use Redis in memory full page cache and your flying .. =)

Just my 2 cents to this thread.

 

[litemere]

This is the second time this week when someone says "it's not a technical article" when I dive into a disagreement with them.

Well, I guess it's a risk you must take - someone will want to go technical if you claim something and they disagree with you.

Yes, you are right that overall performance will improve. I like Litespeed, no doubt there.

And no, even with shared hosting - its important to know how much CPU or RAM your account gets, no need to get dedicated if you server a couple hundreds of visitors in the day but these metrics become important if your application is something heavier than Wordpress site, e.g. - a forum engine.

My main questions to the hosting provider would be in this specific order:

- What backup and disaster recovery policy do you have?
- Where your backup copies are located, same DC or offshore?
- Do you use Cloudlinux or how do you achive user space virtualization?
- Is technical support available 24/7?
- Location of servers if I have geo-based visitors.
- What IOPS limits do you enforce per account and how much CPU gets shared?
- SSD or HDD, perhaps even nvme m.2?
- Flexibility of managing PHP versions, adding or disabling extensions.
- Auto installers for Wordpress and similar so I can drop that Wordpress with a single click.

And I would not ask them about webserver. If I see they use litespeed, I would be like "oh that's cool, + for this one". But it's not even close to being the most important thing.

Soon as we start asking about IOPS, the article is technical.

You bring up valid questions from a person with a technical background. If my mother is looking for hosting, she wants to know will her woocommerce store run fast with their service compared to average speeds among hosts. She is not about to ask about what type of virtualization they use, IOPS limit, the CPU, SSD, HDD, she'll go ABC... If the company is paying for LiteSpeed Webserver, more than likely their overall service is representative of quality overall.

I am going to like your post because overall you make valid and dare I say excellent points if this were a technical article. I feel like I am in a boxing ring with you and you just will not give in!

 

[lucky.sparks]

I feel like I am in a boxing ring with you and you just will not give in!

that compliments me, thanks man!

p.s. webservers are technical as well. mothers will likely look for price & autoinstallers

 

[MJ]

If the company is paying for LiteSpeed Webserver, more than likely their overall service is representative of quality overall.

This is not fair assumption.
I have seen lot of hosts that are using decade old processor and have Litespeed. I mean there is nothing wrong with that, but they don't advertise about server/hardware, just say that hey we are using Litespeed and cpanel, just following buzzwords.

I had come across host that was using E5-2680, while there were others who are using 5950X and Intel gold series. So if your account has 1 CPU assigned, then intel gold/ ryzen will actually be faster than aged processor.
Geekbench of 2680 is ~600 whereas for ryzen is ~1200, so actually 2X faster.

 

[litemere]

apache vs litespeed would feel a lot faster, yeah, that's for sure, I agree 100%.

litespeed vs nginx, the difference would not be so big.

litespeed advantage is concurrency, it can handle more connections than nginx. and overall it's faster for sure.

but my main point is that its not the most important factor, as you said it yourself - location would benefit even more. I would add CPU speed, SSD disks, and of course - smart maintenance of the web application itself.

Its hard to give an example if I did that already few times, but here it goes again.

If the webserver is located US but most of the visitors are EU, speed will suffer for approx 100-150 additional ms and no webserver will fix that.

If application has db bottleneck and there is unoptimized queries that on each load takes 150 ms, no webserver will fix that.

If website has tons of static resources like javascript files, heavy images and stylesheets - no webserver will help there to gain speed if the overall loading will take multiple seconds anyways. Would it be faster than apache? yes. Would litespeed help and save the day? No.

So my main point is, webserver is the least of things you must check when choosing hosting provider, its nice to have but not even close to be the most important question.

btw, your own website loads slow because of many unnecessary dependencies on your main page, for example - no need to load stripe if its not the checkout page.

Check these results out: https://gtmetrix.com/reports/litemere.org/ogc5oKs2/