Google’s Gary Illyes sheds some light on the Google Search indexing process, revealing how and what makes it to the front lines gets there in the first place.
Gary Illyes is a webmaster trend analyst at the tech conglomerate, but also holds claim to the impossibly cheery title of Chief of Sunshine and Happiness at the company. Having worked in the company for almost a decade now, as he joined in 2011, the company wizard is no stranger to how Google and its corollary products function. Illyes is also habitual of responding to user queries and delving more into the software side of things via online forums, cementing him as the best possible choice to reveal how the enigmatic Search engine works.
In an episode of the Search Off The Record podcast, focusing on language complexities on search based selection, Illyes decides to take some time and explain how search results make it to users. Google’s search indexing is apparently powered by three types of storage. Namely:
1) HDD, which is the cheapest and, accordingly, the slowest
2) SSD, which is even faster but far too costly to manage
3) RAM, which is the most expensive and fastest in delivering results
Using the example of a document, Illyes explains that when building the index, storage is assigned on the basis of how often users will search for a particular result. So, if the document in question is more likely to be sought after by users, the index will assign it a faster type of storage, thus appearing that much more easily for web surfers. It’s a pragmatic, cost-effective solution that ensures user satisfaction with the engine.
Not settling for just a simple explanation, Gary Illyes then goes on to provide examples of how index data is stored. Results that might be searched for every second require something fast, thus having RAM memory dedicated to them. Also further elaborated upon is the fact that the bulk of index data is not stored on RAM or SSDs, instead relying on hard drives as a cheaper alternative. A decision that, again, makes perfect sense. Instead of going all-out on purchasing RAM for an ever-expanding catalogue worth of documentable data, it's much better that faster storage only be reserved for the highest tier of sought-after content. It's rather fascinating, seeing the curtain pulled behind on what is essentially a daily routine for perhaps every soul that has ever wandered the internet.
Read next: Google is testing new layout and design for their Sitelinks on Smartphones
Gary Illyes is a webmaster trend analyst at the tech conglomerate, but also holds claim to the impossibly cheery title of Chief of Sunshine and Happiness at the company. Having worked in the company for almost a decade now, as he joined in 2011, the company wizard is no stranger to how Google and its corollary products function. Illyes is also habitual of responding to user queries and delving more into the software side of things via online forums, cementing him as the best possible choice to reveal how the enigmatic Search engine works.
In an episode of the Search Off The Record podcast, focusing on language complexities on search based selection, Illyes decides to take some time and explain how search results make it to users. Google’s search indexing is apparently powered by three types of storage. Namely:
1) HDD, which is the cheapest and, accordingly, the slowest
2) SSD, which is even faster but far too costly to manage
3) RAM, which is the most expensive and fastest in delivering results
Using the example of a document, Illyes explains that when building the index, storage is assigned on the basis of how often users will search for a particular result. So, if the document in question is more likely to be sought after by users, the index will assign it a faster type of storage, thus appearing that much more easily for web surfers. It’s a pragmatic, cost-effective solution that ensures user satisfaction with the engine.
Not settling for just a simple explanation, Gary Illyes then goes on to provide examples of how index data is stored. Results that might be searched for every second require something fast, thus having RAM memory dedicated to them. Also further elaborated upon is the fact that the bulk of index data is not stored on RAM or SSDs, instead relying on hard drives as a cheaper alternative. A decision that, again, makes perfect sense. Instead of going all-out on purchasing RAM for an ever-expanding catalogue worth of documentable data, it's much better that faster storage only be reserved for the highest tier of sought-after content. It's rather fascinating, seeing the curtain pulled behind on what is essentially a daily routine for perhaps every soul that has ever wandered the internet.
Read next: Google is testing new layout and design for their Sitelinks on Smartphones
