The Business Situation
Today, businesses rely on IT infrastructure more than ever before. Datacenters are filling up due to the amount of information required by businesses. Managing the information requires many layers of infrastructure, including servers, storage, and network devices. A recent survey reports that more than 50% of respondents feel they will need to significantly expand their Datacenter capacity sometime in the next 5 years. This real and necessary expectation can have a major cost impact on any business.
The IT Challenge
Infrastructure growth is met with many challenges that can be extremely expensive. With growth comes increased power and cooling requirements, increased administration costs, increased support costs and increased floor space requirements. Each results directly from a diverse server and technology environment.
Left alone, the datacenter becomes filled with racks of servers, each one dedicated to a specific task. Some estimates show the average computing utilization rate in most traditional datacenters is at 15% or less. That’s not a very good return on investment.
Another common problem is rack utilization. Some Enterprise servers or in some cases blade servers, have a power requirement large enough to consume the maximum power a single rack can provide. This means the rest of the rack space is wasted. One solution is to combine servers with high power/cooling requirements with servers that have lower power/cooling requirements. However, this complicates the deployment process, forcing servers belonging to the same architecture to be in different parts of the datacenter. It also complicates support and administration, often leading to confusion for the personnel maintaining the environment. That’s where new technology, smart architectural design and the application of best practices for data consolidation and server management come into play.
The dcVAST Solution
The move to server consolidation begins with a dialogue between dcVAST Solution Architects, the IT staff, and the business managers. This helps to classify the organization’s applications, the infrastructure they run on, and the business drivers associated with them.
We recognize that no two datacenters are identical. Our approach to solving datacenter challenges has five steps: assessing the problem, designing the right solution, implementing the designed solution, supporting the infrastructure, and managing the IT environment.
We start by defining all categories of servers and operating systems, the resource requirements for each application and the different types of architecture (production, development, staging, etc.) for the environment. We carefully work to align the business drivers with priorities such as availability, performance and support levels.
Using this information we can deploy technology with current best practices to save space, reduce power and cooling requirements, lower administration costs and improve utilization and uptime. Some of these techniques include:
This is the technique of running more than one instance of an application on a single server with a single operating system. Web servers are a good example for this approach.
This is a method of running more than one operating system and/or more than one type of operating system on a single server. Some examples include Sun Solaris Containers and LDOMs, EMC VMware and Microsoft Hyper V.
This is a way to make one physical server run as two or more hardware isolated servers. Sun Dynamic System Domains is an example. These domains are fault-isolated, centrally managed and highly redundant. This reduces administration costs, improves uptime and provides flexibility in an environment with rapidly changing resource needs.
Advancements in technology and processor architecture have given us new servers that use less power, less space, require less cooling and in many cases improve performance.
The implementation of one or more of the above techniques can drastically reduce the number of servers required and improve utilization reducing the physical footprint; saving power and cooling requirements; and reducing the support requirements in the datacenter.