Balancing the Loads between Client and Server

If you are shopping for something online, similarly to you shopping at your local departmental store, you want to get the things you want at your own leisure and have them checkout out and paid for at the fastest time possible. If the cashier on the other end is slow in processing your checkout, you may have the tendency get irritated or better yet, just leave without buying anything.

Similarly, if an e-commerce store you went into is slow in its checkout processes, or it process them in an inefficient manner, you will tend not to complete the process and just logout of the website.

Main function of an eCommerce Server

In e-commerce, the essential server or host function is to store, and make
available to a client machine, all of the information that is unique to a store
on the Internet. This unique information will include:

• product data and descriptions
• data controlling the look and feel of the storefront itself
• feedback pertaining to the actual purchases by the customer

In addition to the unique functions in e-commerce, there are numerous generic
functions that must be performed to complete the on-line shopping experience.
Either the server machine or your computer can perform them. Some of these
functions are:

• saving product selections to temporary memory
• computing taxes
• computing shipping costs
• totalling your order
• getting gift wrap instructions
• forms checking
• getting shipping instructions
• remembering who you are and what you bought last time
• generating a purchase order
• sending you an e-mail confirmation of the order
• credit card verification (This is not card authorization which is strictly
  a server side function.)

Your computer, the client machine, will run a web browser application which
allows you to view the unique store information as hypertext pages. These pages
have the capacity to provide rich presentations of audio, video, images, and
text.

It makes no difference if your online store is only serving one customer. The server will not be burdened by the load. But imagine 1 million users at once accessing the server, what will happen? Either the server slows down to a crawl or it just crash.

To put it simply, you need to divide processes with the server and client into equal parts to balance the load.

To demonstrate how this affects you, take a look at a real world paradigm which probably best emulates a virtual world shopping experience. It's the type of store where you view items on a display floor and write down the item number on a pad. You keep doing this for each item you want until you're ready to check out. At which time you go to a central location where a clerk totals the price and tax for everything you've selected; you make payment arrangements, and pick up your purchases.

Now do this in the virtual world using e-commerce software that has to perform all of the generic functions on the server machine. Select an item from the display floor. Write down the item number. Take the slip of paper to the check out counter and ask the clerk to compute your taxes and shipping costs on that item, and hold the slip for you while you look around the store some more. Select another item, take a new slip paper to checkout, and ask the clerk....you get the idea. Server side functionality means that each time you make a selection, you must repeatedly go back to the clerk (server) to hold that selection for you. It also means that if you want to see what your total purchases, shipping costs and taxes are, at any time, you must go back and ask the clerk to make the
computation.

However, the clerk, the server computer, is very fast, so why is this a problem? The problem is that each item you select is about a mile away from the checkout location and the clerk is very busy. You have to walk this mile each time you make a selection. And you have to stand in line waiting while the clerk serves other customers.

How long can that take? Consider that that mile represents the time it takes to request a page and get a response over the Internet...you can judge this for yourself. The clerk is the server machine. If one hundred shoppers show up at the clerk's location at one time....expect to wait. If you're a customer, you may just go home. If you're a merchant, you're in big trouble. You may be losing customers.

The irony of this server side scenario is that while you wait for each transaction from the server, your own computer sits idle. To address this wasted capability, client side storefront software was developed.

A client side storefront runs in the web browser, on your machine. It performs all of the generic on-line shopping functions on the fly. That is: it stores all of your product selections, cost information, and shipping instructions in the browser session; it can save information about you and your purchases on your computer; it totals the order and displays taxes and shipping costs almost instantly; and, it presents a complete purchase order for your review, without delay and without ever having to make another request of the server. In fact the only time client side storefronts have to connect back to the host server machine is to request new information unique to the store or send a summary of
the completed shopping session.

Client side e-commerce software is generally written in hypertext markup language (HTML) using JavaScript scripts to provide the various generic e-commerce functions. Both of these languages are ubiquitous and run in the client machine's web browser without relying on any special server side functions. This means the merchant owns and operates their own store independent of the host computer and any special server e-commerce software. It means that the storefront is portable and can be hosted just about anywhere, from any Internet Service Provider.

If you're a shopper, you want speed. If you're a merchant, you want happy
shoppers. Client side, e-commerce, storefront software delivers both.