We are all prone to wanting more in almost everything life offers. This is also common to buying website hosting – we want more. More hosting space, more processing power, more RAM and more bandwidth as well. We want a bigger, meaner machine which has all the bells and whistles that your money can buy. But we rarely stop to think, whether we really need it and whether that is really the solution to everything. This article tries to reason why you should rethink your purchase of a Dedicated Server for your website or web application.
The most common cause of goofing up a dedicated server decision is due to being misled by your software developers or designers. Many a time, they have no clue how much of the resources they will really need and want to save themselves the trouble of migrating your data at a later stage, hence they coax you into believing that you need a monster configuration server which will magically solve all your problems. The less experienced the developer is, the less empirical or statistical evidence he will show you, for justifying the purchase. 3 years back, we received an email from a friend, asking us for a very high configuration machine. When we realized he was being guided by his developer, we suggested that they do a trial run on a small 5 GB shared hosting account. Till date, they have not needed more than that and are very happy we saved them the huge expenditure.
Before we suggest a Dedicated Server, we request the client to send us an estimate of his requirements and how he has come to that conclusion. Most of the time, the client has made wrong calculations and drastically changed the entire estimation by using predicted figures, which may not actually work in a production environment. This may mean that he has either under-estimated the requirement and will need to upgrade the server in the short term or he has overestimated the requirements and goes for an overkill solution. There are numerous benchmarking tools, which can help predict the exact amount of resources that the website or application requires and also helps test the limit of those resources.
In the past few years we have often heard from atleast 10 different clients, that their future plan is to make an “online portal like Facebook”. While that may seem like a very progressive thought, it is often accompanied by a lack of planning and exaggeration. Ofcourse everyone wants to be the best, but they want to do it overnight, without looking into the path travelled by the industry giants. This over ambition eclipses their decision of doing some realtime tests and then jumping into the project in full swing. We often see that those who take a dedicated server, drop the idea by the 5th or 6th month, when they realize that their requirement was miniscule or their goal is not achievable at the time.
Another common reason why people want to go in for a Dedicated Server is because they don’t want to get down to the root of the issue of high resource usage, but instead would like to work around the issue by pumping in more hardware resources. A blaring example would be that of a small time ecommerce portal, which was so badly coded that the SQL queries kept getting throttled due to being slow, clumsily written and buggy. Rather than modifying the query to its optimum best, the webmaster chose to add more CPU resources by upgrading the server. This solved the problem for a few weeks, till the query started taking more time. Thereafter he went in for a better configuration server. This time, in a few weeks the same problem cropped up and he realised that the problem lay in his SQL coding and not in the server. He has now happily switched to shared hosting and has got the SQL beautifully optimized. His laziness definitely cost him time and money, but it also made him wiser for the long-term.