On Demand in Computing, specially in Cloud Computing this word on demand is frequently used. On Demand has more than its literal meaning. In cloud computing, on demand does not always actually means - if the owner own a certain limit – that is the demand. No, that on demand is basically literal meaning, the phrase on demand in computing has more deeper meaning.
On Demand in Computing : Beyond Cloud Computing
On Demand is not only for Cloud Computing, it is for networking, it is for any science which runs on logic and a pattern of equation can be written. In case of distributed computing, on demand is a general terminology that is used for request served from a server computer to a client computer. It is usually on HTTP GET request. When the user is not using, the demand is lost. Demand lost means – increase in computing resource pool. There can not be any computer or server setup in this World with endless computing resources. On Demand is like an equilibrium with a certain number of users who will always use, it is the absolute number, not specific user; certain number will suddenly increase the work burden. This is the reason, on demand depends on Scalability. Mystically, Scalability and Service Continuity are not equivalent. This is the reason why, most high end research universities, who needs high end computing resources, are using Cloud Infrastructure over distributed computing as in distributed computing the on demand is actually limited too much. On demand is not unlimited in cloud but assured to have virtually endless computing resource.
On Demand in Computing : Examples
The Cloud platform this website is on, has practically endless computing resource. It is impossible to use 30 GB of RAM, 5 load balancers. They are resting as there is no demand. When we will publish this post, there will be a sudden gush of visitors from various social networks, it is configured in a way, the provisioning will also increase proportionately. As Server means hardware, if the hardware behind is crap, the network connections are not optimal, no theory will work.
Another good example of on demand computing is – the technology today’s most mobile broadband uses in HSPDA or higher configuration. If you are not doing anything, like just loaded this website and no other active internet activity, you will notice that, despite your HSPDA can give 7.2 Mbps speed, it will fall back to few kbps. When you will open YouTube, that on demand will work to boost up the speed to few Mbps. They do not always use the phrase on demand, sometimes uses speed throttling synonymously. This is why with Internet Load Balancing, one can virtually have endless dependable speed only bounded by physical hardwares.
Our body has millions of distributed computing system and not all works – fever, pain are on demand as the end response. If I slap you, your adrenaline glad will pour its content on demand in your blood stream and in return, you will get more energy (more glucose in circulating blood, rise in BP) to slap me more hard, the end result what we say – my head got hot. It gets hot on demand to execute an emergency work. Head actually gets hot literally, if you get angry, test by putting your hand over your head, it will feel warm. The energy produced is lost as heat plus there is more circulation of blood on demand. If you get scared, reverse will happen, on demand system by default will fail. How there will be provisioning for escape, that is a separate story.