close

The Impact of Digitalization On Data Center Architectures

Whitepaper SoftwareDefinedDataCenter

Connected cars, smart homes, and an abundance of smart devices and sensors are very much the norm today. But have you ever given any thought to all the data that these devices generate? It is being harnessed and analyzed behind the scenes, to support business decision-making. Hardly surprising then, that such astonishing amounts of data, generated by billions or trillions of connected users, devices and things also needs a new kind of IT infrastructure to manage it: so-called software-defined data centers. We take a closer look at this emerging trend and its impact on a business’ IT infrastructure.

Traditional mainframe environments are known as the ‘first platform’. These were designed to deal with thousands of users and things. We’ve seen the move to the ‘second platform’ of client/server environments, where networks based on x86 servers are now able to handle millions of users or things. Enterprises deploying the second platform have also enhanced their infrastructures with integrated hardware and software based storage and IP LAN or WAN networks. And now, we are now seeing the emergence of a third IT platform. This third platform, a term coined by analysts at research firm IDC, is the enabler of the cloud, big data, enhanced mobility and the integration of social media. It can handle billions or trillions of users and things and is powered by modular, industry-standard servers and new types of hyperscale storage. Finally, the third platform will transform a network into meshed fabrics that are called the “new IP”. Hyperscalability will pre-dominantly be achieved by smart software which can control and aggregate storage media, compute power and network bandwidth.

The emergence of the Software Defined Data Center

This third platform architecture is appearing as a way of managing the growing need to handle unpredictable amounts of data, users, devices and things. It is about creating fast IT to rapidly deliver scalability in transporting, processing and storing vast quantities of data. This requires hardware-independent provisioning of IT resources – and the creation of a software defined datacenter (SDDC).

A complete SDDC comprises several core building blocks:

  • An array of very modular servers which allow the online expansion of processing power. They should also be high density to save space, while efficiency is also important – these must consume only low levels of energy and produce low levels of heat. These servers are virtualized. This is achieved either by hypervisor technology or with a new approach, called containers. You can think of containers as “buckets” of virtual storage. These provide a high level of agility, since users can move workloads easily between different containers, whether these are in their own data centers or external, cloud data centers.
  • A network that enables the movement of data and applications (or containers) across public and private cloud resources and which is accessible by any device from any location. The network should also have software-enabled control, automated QoS management and should have a meshed fabric rather than a more rigid, hierarchical structure.
  • Storage capacity must be hyperscale – in other words, running into the Petabytes – and should have a clustered architecture to provide linear scalability of capacity and performance. It should use open source platforms, and data management should be separated from the underlying hardware using software-defined storage.
  • An orchestration platform is required to provision and manage the server, storage and networking resources. This should ideally also control cloud resources to enable hybrid IT operation.

Before deploying SDDC, businesses have a series of decisions to make regarding which type of SDDC implementation will be the best fit, and this will ultimately depend on their planned usage scenarios. Firstly, they could adopt a functional approach, which is ideal for creating a general purpose infrastructure. This involves implementing discrete, software-defined compute storage and network entities. Alternatively, for some specific scenarios, the SDDC can be implemented using a hyper-converged approach. This method involves deploying software-controlled building blocks, each with separate integrated network, storage and compute functions.

Optimizing Robust IT infrastructures

Although technologies change, the need for data storage is constant. We are certain that these new environments can coexist along with existing, robust IT well into the future. Ultimately, the right architecture depends on the required level of scalability and the applications that will be deployed. Given this longevity, we encourage businesses to optimize their legacy infrastructures.

Storage in particular has great potential for optimization. Many IT organizations have a significant amount of underutilized hardware, as fear of performance bottlenecks often leads to overprovisioning. Today, this presents an attractive consolidation opportunity. We recommend looking at modern storage technologies with strong end-to-end performance architectures, in addition to the ability to automatically prioritize storage performance. Servers also present opportunities for optimization – with the latest modular servers being well placed and easily able to manage both robust and fast IT. In addition, scale-up systems (those that can easily be expanded), especially with large-scale main memory expandability, are perfect for running business analytics applications such as SAP HANA, which run entirely in a system’s main memory.

Transitioning to the third platform and agile software-defined data centers is well underway Scalability is the main driver for most enterprises to start transition. But each business must find its own path, and find the right balance of robust and fast IT. As they plan their roadmaps, even if the full SDDC is not deployed immediately, there are benefits to be reaped by optimizing servers, storage and network architectures.

To find out more about software defined data centers, download our whitepaper: Is software or hardware defining your data center? Or is it your business?

Tags: , , , , , ,

No Comments

Leave a reply

Post your comment
Enter your name
Your e-mail address

Before you submit your comment you must solve the following arithmetic function! * Time limit is exhausted. Please reload CAPTCHA.

Story Page