LiteSpeed compatibility? (drop-in replacement for Apache's web server)

As LiteSpeed is a drop-in replacement for Apache (and can also run along side), is there a possibility to incorporate 100% compatibility with ApisCP?

IMHO, as reliable as Apache is, it simply cannot handle the high-load that some growing (especially uncatchable ecommerce) websites tend to generate. …especially with unexpected traffic spikes.

And while Nginx can out perform Apache on dynamically loading websites, Nginx simply isn’t as smooth of a drop-in for (shared hosting) users who have minimal hosting knowledge, and only minimal experience modifying .htaccess rules. That’s why I feel that LiteSpeed IS the solution of the future. BUT, there’s NO reliable control panel?! CyberPanel is supposed to handle this task, but fails to compare to any notable CPs.

So, if ApisCP could help bridge this gap… something tells me that you might have a lot more customers come knocking at your door. You have an amazing product (from what I’ve gathered) and LiteSpeed Ent is an amazing product. In theory, ApisCP is most likely already compatible (due to the nature of LSWS’s Apache drop-in/replacement)… so, that’s why I’m taking the time to write this up. Hopefully someone’s reading/listening :wink:

For anyone who’s not familiar with LiteSpeed, here’s a quick blog post to read:

And for anyone interested in seeing benchmarks that demonstrate how much more efficient LiteSpeed is (compared to Apache & Nginx), here’s a page w/ screenshots & info:

Worth noting, I haven’t had a chance to install ApisCP on my own hardware yet… but I have purchased a license. I’d like to explore ApisCP more, beyond the demo experience.

I took this step because I need some type of user-friendly and reliable replacement for cPanel… and I’d much rather purchase a license that I might not ever use, as long as it means that I’m supporting independent and free thinkers!

Absolutely not and no intention to vendor-lock to a heavily commercialized product that works in a narrow segment: WordPress and Joomla. Even then, you’re forcing people to upgrade to its commercial variant, Litespeed Enterprise, to get a usable multi-tenant setup.

I would recommend benchmarking ApisCP independently on your own. There’s a guide to walk you through the benchmark process. If you want Apache to benchmark like NGINX and Litespeed, then it must be configured as such. This includes migrating .htaccess directives to runtime directives as well as trimming the superfluous directives that bloat .htaccess rules. This guide will walk you through those steps.

People overprescribe and under-research quite often, no matter the software. WordPress is especially problematic that so much mysticism surrounds proper configuration. Proper configuration is what separates a well-oiled machine from a nasty CVE. Litespeed is marketed as a silver bullet to WordPress but in fact taking a step back to see why these problems exist is a bit more prudent.

From what customers have independently discovered, ApisCP performs as close or better to LS in fact I’ve even had some converts sell their lifetime Litespeed licenses. Most of this lore comes from comparing with cPanel that still uses Prefork MPM in 2022. There’s no reason to use Prefork past 2010 unless you’re floating $540 million in debt. cPanel’s price hikes seem to corroborate they need all the revenue streams they can get.

Plus, you can do some cool things with Apache including integrated DoS deterrence, malware scrubs, any-version Ruby/Node/Python, and rendering optimization through Pagespeed that Litespeed has dropped support for because it conflicts with their commercial endeavors.

Some folks have tried porting Litespeed but the mod_rewrite implementation isn’t truly drop-in. We synthesize addon and subdomains by leveraging filesystem caches to reduce static memory allocation. It results in a lower footprint than your typical implementation.

Definitely benchmark to get a fair comparison. Any statistic can be massaged to conform to any bias you so desire to market.

It’s a touch bias too to characterize Apache as sluggish without giving its implementation in ApisCP a chance, right? :slight_smile:

…but, much of what you’re saying simply is not accurate.

Are their caching plugins limited to the larger ecommerce & CMS platforms? …yes, but they more than likely have a LSCache plugin that comes in your flavour.

LiteSpeed internally offers plugins for:
Magento, WordPress w/ WooCommerce, Joomla! w/ VirtueMart, XenForo, PrestaShop, MediaWiki, Drupal, Craft CMS, OpenCart, Laravel

BUT, what about those folks who aren’t using one of the one’s listed above? Well, LiteSpeed server also optimizes websites straight out of the box… without any additional plugins needed.

Maybe I’m not truly following, but IMO, there’s no “vendor locking”.

  1. Apache Web Server can run right along side LSWS (LiteSpeed Web Server).

  2. anyone can use/run LSWS single tenant for FREE. in fact, I’ve taken a lot of criticism (by friends in “the know”) for purchasing a LiteSpeed Ent license. they all highly recommend creating single-node Docker containers for users, boosting security while minimizing cost.

  3. “Silver Bullet”? …well, when it comes to WooCommerce, I’d have to agree with that. I’ve yet to find a single caching plugin that can efficiently cache Woo’s unique customer fragments. It sounds like you’re saying that ApisCP runs so smoothly that there isn’t a need to worry about page caching… which is something I’ve yet to experience, simply because I haven’t developed a WP website on ApisCP yet. (But if you’re saying it’s that good, I do have a resource hog of a WP/Woo store that I’d love to throw at it.)

  4. Some folks have a strong need for the Silver Bullet. While some of us can figure out where the problems are generated from, much of the hosting population doesn’t even know what FTP is. The typically shared hosting clients needs help in this area. So, unless you’re saying that (again) ApisCP handles WordPress’s shortcomings with ease, making it run “properly”… I don’t want to be the type of reseller that has to assist each and every one of my clients with their websites, when a simple “activate” and “1-click optimize” is all that they’d need to do. …I’d prefer to focus on growth, and feel confident that my scalability is simply covered. (which it does sound as though you have covered w/ ApisCP)

  5. Google Pagespeed IS built into LiteSpeed Ent.

  6. I NEVER meant to come across as criticizing with my post on any level. While I did sound bias to LiteSpeed over Apache, that’s purely due to how much better my LiteSpeed server is running an account I migrated from cPanel… straight out of the box. I had mpm_event, CloudLinux w/ lsphp, and CloudFlare Railgun serving the website on my cPanel server. On LiteSpeed, the same page loads within a second. I realize that you’re saying part of the problem is coming from the nature of how cPanel works… but we can’t blame cPanel for everything (as much as we’d like to).

…so, in review… I am looking forward to giving ApisCP a run for its money. (Possibly even a bit more so now :wink:). Worst case scenario, I simply add LSWS Ent’s ‘Free Starter’ as a prebuilt Docker Container. Sounds like I might get the best of both worlds that way, right?

Avoid jumping to conclusions without taking the panel for a ride first.

IMHO there is literally no reason for LiteSpeed to exist in the ApisCP stack, there’s no literal proof it performs better when compared to ApisCP-optimized Apache. Seems to be just bias as Matt says towards a not-quite-well-configured Apache past experience that people have, bridged by a switch to Nginx or LS that brought a tiny bit more speed or literally nothing else (placebo? lol). Keep in mind that most people don’t just switch the web server, but a whole hosting server and that means that you can get on a new server that probably has faster CPU, faster RAM and faster drives or super-fast drives. That all counts towards the perceived speed, and may not be all thanks to LiteSpeed…

PageSpeed is built into ApisCP’s Apache version too, did you have a read of the specs and infrastructure or is this just 'cause you heard Apache buzzword instead of LiteSpeed? Give ApisCP a ride first, then we can chat more about this.

I also can provide a real-world example, but that won’t mean much if you keep your bias for a bad past experience instead of testing the actual thing on the field.

Stack convolution isn’t a design goal. If you’re running another HTTP server then what you have isn’t adequately addressing the requirements.

Yes and there’s no request-time pickup of .htaccess directives, which makes it a poor fit for multi-tenant environments.

Stack convolution isn’t a design goal as noted.

Page caching does exist in Apache, both memory and disk-backed. will walk you through the steps where we start to find a reasonably fair baseline for Apache to NGINX to LiteSpeed.

Yup, that’s the point of ApisCP. We bring all facets under a common API without relying on third-party support. will walk you through benchmarking without subscribing to third-party solutions.

Web > Site Optimizer > Safe optimizations. Can’t miss it. Included in ApisCP at no cost.

Yes, this is in line with:

Litespeed has dropped support for because it conflicts with their commercial endeavors.

Ultimately I’d encourage you to see how ApisCP fits your needs by benchmarking it first. This isn’t the first time I’ve heard about LiteSpeed nor do I anticipate it to be the last. LiteSpeed has no use in ApisCP’s stack nor ethos of being able to standalone, unencumbered by third-party licenses to just work.

If you have any other reservations then I’d encourage you to drop by our community Discord. There are clients on there who have switched off LiteSpeed, pared back monthly expenditures, and never bothered to look back

Technology is always moving, so I certainly hope you’ll give this an honest look first :+1:

1 Like

Last august we had a sale go nuts on a WooCommerce shop hosted on ApisCP. There were some hiccups indeed, but none regarding the web server itself. The only things we had to tweak were MySQL concurrent connections limit and PHP-FPM workers. The shop handled 2-days of continuous orders topping at about 1M in revenue. Pretty impressive let me say—totally wouldn’t be able to do so in my previous Plesk setup.

The test doesn’t have any meaningful data. You can’t test three different environments and simply compare them to determine the best one. You’d want for your test matrix to be all three plugins tested across all three web servers in order to have truthful results.

Please mind that tests may vary even if you switch from packaged to compiled software too. Packages are built to appeal the general public, while compiling the same software on your own machine may introduce specific optimizations based on your own machine.

There’s simply too much stuff to test and no one wants to do that when testing specific configurations plays in your favour.

Let me state this again… I’m not trying to speak down on Apache! I was simply trying to ask a question, as a search for LiteSpeed resulted in zero results.

And my performance experience that I speak of is literally the same server architecture, reformatted and reinstalled.

I’ll revisit this thread after I give this panel a go.

1 Like