So lets have the little discussion on this new InnoDB feature design concept. There are 4 important design concepts below

1. Inverted Index
2. Full Text Search Index Table
3. Full Text Search Index Cache
4. Full Text Search Document ID

Inverted Index

Like all full text search engine InnoDB full text search engine have inverted index which separates the incoming text into words and store them in auxiliary tables. For every word the document ID list store which is called document ID position pair list. So there are two columns in inverted index tables are word and ilist. As we are storing the word position against of each word so Proximity Search is not possible which is the drawback of MyISAM.

Index Tables

Index tables also called Auxiliary tables so the words in the inverted index tables stores in the auxiliary table and also called Full text search index tables. These tables are separated as partitioned so the words are partitioned into 6 tables against there document ID’s.  User also can define the partition scheme manually according to the charsets.

Index Cache

Full text Index are referred by a set of index Auxiliary Tables on the hard disk. There also exist a full text searching index cache to execute the results before flush to the Index table. The Index Cache is the black tree that stores the word document ID/position ilist information in it. There is a difference between this cache and index tree which is only used to cache the word inserted. So not every time words in the queue called from the disk for cache. In innodb the index table updates regularly.

Document ID

The most important concept which will change the Document ID management, every word have a unique Document ID represented by “FTS_DOC_ID” automatically added to the user table. Also there is the unique index on that column “FTS_DOC_ID_INDEX”.

How to Use INNODB Fulltext Index

Example of creating a table and unique index

Create a table

CREATE TABLE fts_test (

FTS_DOC_ID BIGINT UNSIGNED AUTO_INCREMENT NOT NULL,

title VARCHAR(200),

body TEXT

) ENGINE=InnoDB;

Create a index

2) CREATE UNIQUE INDEX FTS_DOC_ID_INDEX on fts_test(FTS_DOC_ID);

3) Load data

4) CREATE FULLTEXT INDEX idx on fts_test (title, body);

Benchmarks Of Sorting Data

There is new system for controlling the parallel sort degree “innodb_ft_sort_pll_degree” new variable range from (2 to 32 max). This is used to how many ways you want to parallelize sort operations. So this is an Experiment results using innodb_ft_sort_pll_degree different values

Following table shows a quick run on 2.7 G wikipedia data on a 8 core machine:

The test (http://www.webpagetest.org/result/110518_8Y_c7e2adb4e09ebe56465372560230d115/) of web page test is showing that sprite of bottom logos is taking up all the page.

Best thing which I liked about IE 9 is the feature which tells which add-ons made IE to start slow. I only had few add-ons for IE as compared to FF. On start up IE showed following screen at bottom of screen, notifying about slow add-ons.

Upon clicking “Choose Add-ons” showed up following screen.

This showed the culprit add-on, responsible for slower startup. I immediately disabled “HTTP watch”, because I was not actively using it. Few other add-ons were shown, from which I disabled most of it as I was not using those.

I found this feature very helpful for better experience. This allowed me to get rid of many add-ons which I really not using and just paying cost in term of CPU and memory. FireFox should also introduce similar feature, as more and more add-ons are making FF slower to start.

To review add-ons performance in IE 9 you can also go to “Tools > Manage Add-ons”, where you can enable/disable any add-ons.

For website owners it is recommended that they install IE 9 and test their websites compatibility and performance with IE 9, because its usage is expected to get a spike as its being rolled out to different regions.

Stats already showing increasing usage of IE 9 in Europe, but still its way behind FF4. In Europe FF is still managing to be at top in usage.

Source: StatCounter Global Stats

First I tried to google for download link of IE9. If you see screen shot, first link (paid link) took me to Arabic version of download page, where every thing was in Arabic. If you note 2nd link is for beta version of IE9 so it is of no use.

Finally on 3rd try i could find the download page. The download failed 3 times, before i was able to download. As this installer was online installer, I started searching for offline installer of IE 9. The download completed in reasonable time.

Although i downloaded an offline installer but installation screen showed me “Downloading” screen. The “Download” bar stuck at same place for more than 10 minutes.

Finally the installation completed and as per “Microsoft Policy of Annoying” it showed screen to restart my PC before I can use IE9.

Over all my experience of upgrading to IE9 is very bad, and don’t have much hope of good experience of IE9 its self.

My point is to encourage developers who get too deep and think about the best practices, best architecture etc etc, to write something which is beneficial. If it is a success you can always have best people in design, architecture, scalability, bla and bla to help you. So start writing what you think is useful without concerning much about too many things at the moment.

1. Serving images from cookie less domain
2. Web Performance Optimization
3. Magento getting timeout errors, even with max_execution_time set to zero, on WAMP

To solve this issue you need to do following steps

1. Download this .pem file from http://curl.haxx.se/docs/caextract.html
2. Place this file on you web server, where you can refer it in code.
3. Now add following code where you are doing the actual cURL request curl_setopt($curl, CURLOPT_CAINFO, “c:/path_to/cacert.pem”);

Please ensure that you provide correct physical path to cacert.pem file. A sample code will look like as given below
// the url to connect, note its HTTPS url
$the_url = "https://www.google.com/";
$certificate = "c:/path_to/cacert.pem";
$curl = curl_init();
curl_setopt($curl, CURLOPT_URL, $the_url);
curl_setopt($curl, CURLOPT_CAINFO, $certificate);
// adjust these as per your need
curl_setopt($curl, CURLOPT_HEADER, 0);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
$response = curl_exec ($curl);
curl_close($curl);
echo "the cURL response is: ".$response;

This should resolve the error.