The landscape of web servers is rich and varied, with technologies that constantly evolve to meet the increasingly sophisticated needs of users. Amidst this cacophony of solutions, LiteSpeed stood out, gaining ground and grabbing industry attention. Many see LiteSpeed as the future, a panacea for the needs of web projects, from the smallest to the largest. However, a deeper and more detailed reflection may reveal a different reality: LiteSpeed's popularity can be attributed less to its inherent technological superiority and more to a widespread lack of technical skills among systems engineers.
NGINX and Varnish represent the cream of web server technology. Used by top-tier enterprises, gigantic media organizations, and e-commerce platforms with huge volumes of traffic, these tools have become synonymous with superior quality and robust performance. Their open source architecture and broad developer community support ensure unrivaled flexibility, making them highly customizable tools to meet the most diverse needs.
We talked about it extensively in this post LiteSpeed Web server review. However, it is really difficult to propose a better technology, when the public hears everywhere about alternative technologies, more famous and portrayed as the best solution on the market. Exactly what happened in the VHS VS Betamax story is happening.
In the 70s and 80s, Betamax and VHS were the two major videotape formats, creating a famous "format war". Sony first introduced Betamax in 1975, with much higher recording quality and in less space. However, JVC introduced the VHS format the following year, which allowed for a longer recording time. Despite Betamax's superior quality, VHS became more popular due to its longer shelf life, and the support of more manufacturers. Eventually, VHS emerged as the de facto standard in the home video industry.
However, the power of these tools lies in using them correctly. It's not enough to simply install them on the server and leave them alone. Each technology requires careful management and customized configuration to be used to the best of its capabilities. And this is where the role of the system administrator becomes crucial.
A competent systems engineer understands the intricate details of server stack management. It includes how to adjust the cache, how to check the Cache Control and how to perform a selective cleaning of the cache. Has the expertise to handle performance and hit ratio issues. He understands the potential of NGINX and Varnish and knows how to leverage them to provide an unmatched user experience and optimal search engine rankings.
Unfortunately, the industry is full of systems engineers who lack these crucial skills. This skill shortage can lead to sub-optimal implementation and management of NGINX and Varnish, which in turn can lead to performance issues and a disappointing user experience. Unsurprisingly, in these scenarios, LiteSpeed may look like a more attractive and promising solution.
For example, a particularly significant challenge is selecting an appropriate cache Time To Live (TTL). A proper TTL for caching is critical to ensure that cached content doesn't become stale, degrading the user experience. Unfortunately, many systems engineers aren't familiar with the nuances of this critical setting and may choose a TTL that isn't adequate for their client's specific website needs.
A related issue is cache selective purge handling. This is an essential technique that allows you to selectively remove specific content from the cache without having to flush the entire cache. This is especially important for large and dynamic websites, where clearing the cache altogether could lead to significant performance penalties. Again, many sysadmins don't have the expertise to implement this feature properly.
Finally, there is the matter of optimizing the hit ratio. A high hit ratio means that most requests can be served directly from the cache, thus improving performance. However, achieving a high hit ratio requires careful configuration and monitoring, skills often lacking in less experienced sysadmins.
In these following Screenshots we can see two Hosting managers, the first a professional Hoster complete with SRL, the second a semi-professional Hoster who however is preparing to try to obtain the assignment of important orders. It is self-evident and undisputed that their considerations are absolutely wrong across the board.
In the above case, for example, the gross error is to think that every time the user or the editorial staff writes a post, the cache must be manually cleaned or even bothered by the system administrator in doing so, when it is clear that there are plugins capable of hooking hooks such as WordPress save_post() and proceed with a selective PURGE of the post, categories, feeds and any sitemaps. While underlining the seriousness of this statement, it is fair to consider that the fact occurred a few years ago; therefore, it is now certain that this misunderstanding has been overcome on a logical level; however, it is also true that this supplier has never offered enterprise services based on NGINX and Varnish but has promoted itself by directly proposing LiteSpeed in order to find a more suitable service for the skills. As a practice, we do not mention names to protect and respect the professionalism of each supplier.
As can be seen from the screenshot above, we are faced with an example of a manager who, while offering hosting services, does not fully understand the advantages associated with a long TTL.
A long TTL, for example of a week or even a month, can offer significant advantages, especially when dealing with content that does not change frequently.
However, setting a TTL as low as 12 hours, as in this case, can create a number of inefficiencies. One of these concerns the speed with which content is served to users. With a shorter TTL, older content is evicted from the cache more frequently, meaning it has to be fetched anew from the server each time. This process can increase server response time, known as Time To First Byte (TTFB), especially if your site's backend or application is not optimized to provide high performance.
A longer TTL, on the other hand, allows content to be held in the cache for longer periods of time, reducing how often it has to be fetched from the server. This can lead to a significant improvement in TTFB and consequently a smoother and faster user experience.
Additionally, a longer TTL can reduce the load on the backend server, saving resources that can be used to manage other critical aspects of the site. This can be especially beneficial for sites with a high volume of traffic or limited server resources.
The fact of finding also in this case absolutely an exception to speak of selective Purge when speaking of Varnish Cache, suggests that there are big gaps to be filled and an absolutely insufficient cultural baggage to offer hosting services at a professional level.
Obviously these are just two obvious examples of how much improvisation can reign in this world of "professionals" many of whom also lack academic training, such as a degree course or at least related studies. In reality, the problem is much broader and the gaps are truly many and varied.
That's why Litespeed while having less flexibility in customization and implementation of complex projects, however has to its advantage a basic configuration that is sufficiently better than most of what many "colleagues" are able to offer using more powerful and complex tools such as NGINX and Varnish.
However, this perception does not imply LiteSpeed's superiority over NGINX or Varnish. Rather, it highlights the importance of system skills for the correct management and configuration of the server stack. LiteSpeed's success is therefore a reflection of the shortage of systems engineering expertise in the industry, rather than a tribute to its technological superiority.
Similarly, companies like WordPress VIP and Fastly, which handle hosting for big-name clients like CNN, The New York Times, BuzzFeed, and many more, demonstrate the effectiveness of NGINX and Varnish. These companies have invested heavily in developing the technical skills needed to manage and configure NGINX and Varnish to their full potential. Unsurprisingly, they opted for these technologies rather than LiteSpeed.
In conclusion, it is important to remember that the choice of technology should not depend on its popularity, but on its effectiveness in fulfilling project requirements. While LiteSpeed may seem attractive for its apparent merits, NGINX and Varnish, when managed properly, offer superior performance, scalability, and flexibility. What really matters is the systemic expertise needed to make the most of these tools.