Transformation From Traditional Infrastructure To Cloud Infrastructure

Nowadays most organizations have replaced their traditional infrastructure with a cloud infrastructure. Do you know why? We will discuss the basic concept of how the traditional approach is transformed into cloud infrastructure in this article. First we will see the transformation of traditional approach to a virtualization approach, then virtualization approach to cloud approach.

Traditional Infrastructure

Evaluation From Traditional Infrastructure To Cloud Infrastructure
Before the early 1960s a dedicated physical server was used for each application as shown in the above picture. A physical server is a hardware on which a server operating system like Windows or Linux is loaded, then on top of the operating system we can run the needed application. The only pros of this approach is there is no resource starvation between the applications because of the dedicated hardware.
Physical server cons:
  • No resource sharing- Consider Server 1, Server 2 and Server 3 has 99%, 5%, 45% utilization at a time respectively. Server 1 needs more resources but Server 2 and Server 3 have more resources. Even if there is resource vacancy server 1 cannot use it. So there is resource wastage.
  • Hardware utilization is very low - CPU and OS have their own load average. Normally in typical system load average is greater than the OS load average. Even the CPU can run in high frequency. If OS is not that much of capability then there will be a low utilization of the hardware.
  • No high availability- If a server fails the service provided by that server also has more downtime. For adding memory server should be shut down
  • No cost effectiveness
  • Harder to manage
So, there are many cons in a traditional approach. So the solution for this is virtualization technology.

Virtual infrastructure

Evaluation From Traditional Infrastructure To Cloud Infrastructure
Simply put, Virtualization means Multiple Operating Systems on a single physical server. That is, we convert the physical resources into virtual resources and then these virtual resources are divided into multiple virtual machines (VMs). Each VM runs its own operating systems and has their own virtual resources and behaves like a traditional server.
One physical server has its own physical resources like CPU, Memory, Network Interface Card, and Hard Disk. After the virtualization each VM has its own virtual resources like virtual CPU, virtual memory, virtual Network Interface Card, and virtual hard disk. On top of this virtual hardware you can load a guest OS and then you will have your traditional server.
In the above picture you can clearly see that one physical server is virtualized out into three virtual systems. So there are three systems on a physical server. So this virtualization technology evolves traditional infrastructure to virtual infrastructure.
  • Increasing the utilization of the hardware
  • Cost Effectiveness
  • The total virtual resources can exceed the physical resources –using dynamic resource allocation
  • Time to Market- Can create a new VM easily and provide needed service.
  • Easier management than the traditional infrastructure
  • No down time for performing resource upgrades
  • Single point failure with regards to the Hardware- That is, if the physical server fails all the services will be down
  • Not fault tolerant
  • Security Holes - Malicious programs can make the resource starvation between the OS
  • It will pool the resources only in one Server. More than one server cannot be pooled together
Then the solution for this is Cloud Infrastructure

Cloud Infrastructure

Evaluation From Traditional Infrastructure To Cloud Infrastructure
As shown in the above picture multiple physical resources are pooled together by a cloud platform. All are integrated as virtual resources. Then the VMs can be created with their respective virtual resources.
  • Full resource utilization
  • Cost effectiveness: Don’t need to spend a huge amount to purchase the equipment.
  • Time to market is low
  • 100% fault tolerance- Even if one hardware fails the service will not go down.