In our experience, every day we encounter dozens of companies who waste their computational power, their server to go to really provide a few thousand visits, to a very high cost. Legacy of the past, the use of dedicated servers is often not optimized on servers with different cores and GB of RAM, you are finally able to serve a few hundred visits , below is the real potential hardware.

The reason: These servers are not well configured, do not have any type of optimization, and often result in damage to the company where they slowly serve web pages and the hosted websites

Recycle your old server, switching to VPS Hosting and instances of Cloud Hosting

There are a large number of dedicated servers or colocation who have several years of service and have a very high operating cost for their clients. We’re talking about machines that have processors with minimum 2 Core , 2 to 8 GB of RAM , but in fact mount older motherboards and have a high consumption compared to newer servers.

These servers should be replaced, and the best choice, to date, almost always is to use instances of cloud computing, perhaps provided by the same provider from whom you purchased the dedicated server. The substitution instances of cloud computing does not necessarily mean spending more, in fact, very often the server we are going to replace, being physical, are redundant with other machines, increasing the power consumption and cost: this need with the demands of cloud computing is not always necessary, since many vendors offer HAS base on virtual instances and the lack of having RAID or other common systems on the physical servers to increase the reliability of the system.

Matching Resources And Performance: Optimize Server Daemons

Your server does not necessarily need to be replicated in the cloud with the same resources, we have already said that the main problem that is encountered is the lack of service configurations within the server and the lack of a general optimization of the machine. This leads to a high consumption of resources, not justified.

What Are The Services That Most Affect Performance:

Webserver : Most have a view of the servers have an installation of Apache Web server, but this is not the best solution for all your needs. Apache has become one of the most popular, but nowadays there are better solutions such as Nginx and LiteSpeed, solutions that are configured for reception of high traffic loads.

Database : In this case we know that the problem is almost always MySQL. It is a database that does not scale quickly and which requires a configuration for ad-hoc environments which may vary according to the type of site. In these cases, our advice is to proceed first to identify what is the actual load on the DBMS, performing analysis using specific tools and proceed to the configuration file of MySQL: use successfully for some customers even Percona MySQL , tools that we repeatedly reviewed and for which there are several applications which are already ready for configurations.

The database is definitely one of the most delicate and critical for those involved to migrate its infrastructure with a performance optimization. At present we found that the greater improvement is achieved initially by an ad hoc configuration of machine and balancing resources properly if the machine takes care of serving only the DBMS, and going to provide a configuration tailored to the individual service installed, precisely with the help of a prior analysis of the problems. Very often the database is a “victim” of queries in poorly designed applications, a problem which can be solved easily without putting hand directly to the application code.

Load Testing And Performance Testing : Once a server is put together, the job is not done, we can even say that this is only the first step. What is needed is to see how does the server and its configuration under “stress”, going to simulate a load as usually we do in our reviews. This allows you to see if the work done during the configuration is correct, and especially how it responds to our server in the event of a high peak of requests: usually both hardware and software configurations must provide the level of “tolerance” that help the system administrator to exploit the resources for the management at the peaks.