Researchers recently released a pre-print (article sent to a scientific journal) about the SIEVE algorithm, which promises to make internet browsing much faster. The invention can reformulate the current cache system, which records small files from websites to make access to websites faster.
Presented in December 2023, the article talking about SIEVE explains that the new system proposes a way of reorganizing caches to make the website experience more agile.
Site caches will stay on devices for optimized loading (Image: Getty Images)
“The main reason computers and the Internet are fast is caching. We feel that software caches are this ubiquitous but underappreciated pillar that allows the modern web to function, and so working on them can have an outsized impact” , defended Yazhuo Zhangwho is a doctoral student at Emory University in Atlanta and co-author of the SIEVE research, in an interview with the website Live Science.
The full paper on the discovery will be presented at the 21st USENIX Symposium on Networked Systems Design and Implementation, which takes place in April this year.
How does the SIEVE algorithm work?
The SIEVE algorithm has systems such as FIFO (First-in, first-out), which adds objects on a sort of priority mat. When the objects reach the end, they are discarded.
Another tool used by the new algorithm is the LRU (Less Recently Used), which operates in a similar way to FIFO. However, in the case of LRU, when data from a website is requested by a browser, it goes back to the beginning of the queue.
In practice, the SIEVE works from the binary model, choosing site elements on a “priority mat”. Site elements are first characterized as “zero” upon first access and then as “one” when re-accessed.
SIEVE promises to revolutionize the experience of accessing the internet (Image: 07_av/Getty Images)
That said, caching websites used more than once is easier to request. However, when an object is characterized as “zero”, it leaves the queue so as not to affect the site's loading.
The creators also emphasize that the possibility of bugs is also lower, as the new algorithm was created with less than 20 lines of code.
When will SIEVE be available?
Initial data shows that the page loading algorithm has been implemented in 10 libraries that house famous websites. With this in mind, the Meta and Google are studying ways to put this in their web products.
Big companies are keeping an eye on new algorithmSource: GettyImages
Initial tests showed that when put to the test against the nine most famous algorithms, SIEVE outperformed in 45% of analyses.