Real time pagerank?

I keep hearing people talking about Google covert real time page ranking, that we are just seeing the toolbar update on a whim. Why do you all think this? After working with several very large projects using massive clusters I just dont think its possible.

Everyone seems to agree that Google assigns a viewable page rank to a website or page every 80-120 days. Depending on when you publish your page, this means you may have to wait anywhere from 1 to 120 days to find out what its page rank is. What is much debated though is how often Google actually recalculates page rank. For some reason lots of people are under the impression that at any any given time Google knows your exact current page rank. But if you look at the data you will understand that real time page rank is not possible. There are countless reasons why it can never be real time or even achieved with only marginal latency.

First lets assume there is no latency between Goggle's data centers, and they have arrays that can crunch the numbers in a few seconds/minutes. Now you publish a page at 1.com/a.htm. Google sees the new page and domain and gives you a 0PR. 5 minutes later 2.com/b.htm which carries a PR5 backlinks to your site. How long before Google sees the backlink and correctly updates your PR? To further complicate the issue what if 2.com/b.htm picks up 10000 relevant backlinks and jumps to PR6? The pagerank formula is too dynamic to afford real time page rank assignment. Snapshots are required which would make the data static before the formula can be run.

Second look modern hardware limitations. Even if every data center was running arrays of IBM Blue Gene super computers writting to U640s drives there would be too much information and resource demand for real time page rank. While Google maybe able to index sites in only a few minutes this is achieved over a massive array of servers scattered across the globe. The simple fact there are 30,000,000,000+ web pages and only so many FLOPS of processing power available shows that a PR update requires anywhere from several hours to weeks.

If you watch the PR updates closely you will see changes happening over several weeks. Backlink updates, page ranks resetting, data centers showing different results. To say that Google is capable of real time updates is saying that the whole two week deal with back links, data center variations, and so on are just smoke and mirrors. Running an array at load vs idle is a huge monetary difference. No corporation is going to do it for the sake of smoke and mirrors.

All of this is not saying that Google does not do some of this to limit analysis of the algorithm. If Google could do real time page rank updates I doubt they would as many others have said. It would lead to abuse and make the system too predictable. In short just because you can does not mean you should. But based on current technology and architecture real time page rank updates are not possible.

Thoughts?

 

it must be imposible to have a live pagerank, even a dayly updated pagerank its impossible, just imagine how many sites there are on the internet.

 

I believe they have daily updates, wich occurs once during 24 hours for all sites at different times.

"Real time" is not possible and is not really needed, at least google have fast indexing...

 

google algo is based on the page rank system.

in my opinion, they are updating pr in shorter time periods for google search but they are updating toolbar pr 4 times in a year

 

google toolbar should give more accurate PR . I don't trust other sites.

 

Then why does the pagerank update happen over several weeks as far as population and so on? Hell DNS population is not ever real time. Even if one datacenter has the current PR for a site, how long is it until the data center I use would have the information? Since you are saying its real time I guess it would be the other way around. So if a spider from one data center finds a back link, how long before it reports it to whatever data center calculates page rank. There are days of latency involved. Not the 3 months the tool bar reflects but its not "real time"

 

People keep saying it's a 4 times a year export like that's biblical. It's already August 9th, so I don't think we'll be seeing 2 more updates this year.

 

I tend to think the same.

May be they are going to 3 or may be 2 (looks like they are getting slow)
And they are taking too long once they start the process.

Many errors and corrections as well

 

"Give me Live Page Rank or give me death!"

 

If you join the google sitemaps / webmaster tools you can see the live pagerank of your site in aggregate form (shows you the number of pages from your site that are unnassigned / low / medium / high PR). For my newer sites I have several medium - but my toolbar pagerank is still 0 and will be until the next update. Meanwhile my site is getting indexed at a PR5+ rate and receiving traffic on 3rd level pages.

The PR system is live. If you get a link on a page that is frequently updated it will effect you quicker. I have live traffic analysis on my sites (of my own design). I remember the first PR8 backlink one of my sites received. Suddenly my site was being spidered like crazy and I went from around 70 indexed pages to thousands. All of those indexed pages lend to the keyword relevance of my internal pages - so SER went up steadily for a few months. I think it takes a little short of two months for the effect to propogate.

People that say PR has "nothing" to do with search engine rankings have only scratched the surface of the art of SEO.

Some sites get benefit more than others. It depends on site design.

 

i guess they are kidding who talks about real time page rank...

 

I think the poster is projecting the context that we speak of "live" pagerank with the context that we speak of "real-time" pagerank.

It seems that Google spreads the PR juice every time it crawls a page - thats why its called live. Some people might say real-time when they mean live. Obviously nothing is real time. Search engine ranks do not typically change overnight either - but that is mainly because of the intermingling of off-site and on-site factors.

Snapshots are not required - because the PR is just a dynamic factor in a larger gaming algorithm. Faults are tolerable. The Pr spreads at the rate of crawling. The rate that pages are crawled depends largely on PR and how frequently your content changes - although a low PR site with frequently changing content will still be crawled infrequently. I still see PR 4 and 5 pages where the last time it was crawled was months ago. This might mean that the page has lost alot of PR - but it also might mean that google has found that the pages are rarely updated.

 

The Census Bureau is always collecting data. That does not mean thet have reports and statictics on hand. I am not saying Google only checks the week once a month or so. I am saying anyone who has worked large arrays know their limitations.

Or the pages have something like <META NAME="REVISIT-AFTER" CONTENT="60 days"> or the interval is set in the sitemap. I have several PR 5/6 pages that are set to revisit once a month or so.

 

I think google has the computing power to make it happen