2 questions on HD caching of pages

I've been continuing to dig into the impact of PB on my DNN site and two more (possibly related) questions have come up:

Question 1:

As I expect indivudal blog enties ('entryid') are being cached in a folder, one folder created for each entryid. However I note that some folders created for an individual 'entryid' contain more than one cached file, some of these folders - for example entryid 744 - have 10 to 12 different cache files, each of these files having a date stamp from the last four days. I then checked in my Google analytics account to see if in the last 4 days more than 10 or so 'URLs' with an entryid of 744 were visited. I found that only 2 different URLs were requested in the last week with an entryid of 744. Now I am wondering who requested the other 8 or so? Someobdy must have as these cached files were created, no? One scenario is that a search engine bot requested the file and this would not have turned up in google analytics (as they only record and report visits by agents that execute javascript, which SE bots do not). Or do you have an explanation as to how so many files were cached for one entryid?

Question 2:

One of the avenues I went down when  looking into the previous question was the calendar function that comes with the DNN blog module (a user can naviagate via the calendar to see, for example, blog entries published in previous months). I don't think this is the source of the problem above (becuase once you click on a link in the calendar the URL you are brought to does not have the enrtryid). Nonetheless another point is brought up by the URL that a user arrives at after clicking a link in the calendar. The URL they get to has a structure similar to the following:

• www.domain.com/dnn/Blog/tabid/81/BlogDate/2008-08-15/DateType/day/Default.aspx ('day' is replaced by 'month' depending on what is clicked)

Considering the additional parameters in this URL (highlighted in bold) is it worthwhile to add for the tabid the following extra 'page parameters' in PB's 'cache settings':

•  BlogDate;DateType;day;month

Or to pose the question another way: if they are not added how would PB deal with these pages when they are requested? Would they be cached? If so in what folder? Is it that if these parameters are not declared that all cached 'calendar' pages would be deleted when a new calendar link is clicked?

As you can see I am still trying to get my head around some of the finer details of how PB works. All clarification is appreciated.

Eoghano

Good questions.

For number one:
My guess would be spiders, or just different browsers. Another part of the cache key (other than the Url) is the value(s) for ext-cache-parameter. These are PB Tokens and can be cookies, form values, etc. By default the ext-cache-parameter is set to allow at least browsers that do not support javascript to be cached differently.

For number two:
The page-param is directly related to the cache folders. There will be a new folder created for each page-param.  What this does is allow you to segment up pages that have blogs, forums, articles or other "multi-page" modules on them.  The benefit comes when a "page" is updated.  PageBlaster will remove all cache files in the folder that is updated. So if you had blogs without any extra page parameters, then all pages will be removed if a single post is updated.  This is not a big deal for smaller sites, but it could cause some thrashing on larger sites that get lots of read and update traffic.  On the other hand I don't recommend putting in too many extra parameters because each request has to have all evaluated to see where the cached page goes.  The best thing to do is to add just enough so that it acts like a real "page".  EntryId, ThreadId, ArticleId, are all you probably need.  You don't need to add parameters that are just extra data and would never get a cached page in them because one of the other parameters is always there too. So to answer your question, I would not add anything other than EntryId and BlogDate on the Blog page.