The way data is processed and maybe more importantly the amount of data produced today seemed incomprehensible when computers found their way into offices for business use. As such, data centres reached breaking point as they became inadequately stocked to deal with the significant increase of data. This was, and still is, an issue that is getting harder to deal with as each day passes because more data that needs to be securely stored is churned out on a huge scale.
Remember, in everyday life the amount of information that needs to be waded through to find a specific tweet or a blog post from five years ago is significant and something people do not tend to think about a lot.
Think about your hard drive, you may have 1TB of internal storage available to you directly on your system. Using that as a reference, it may surprise you to learn that in order to access what may seem like the simplest bit of information online, over 10 million GBs or 10,000 TBs of information need to be accessed.
The existing infrastructure
Initially, the infrastructure of data storage was three-tiered. Starting with application moving on to database and ending with storage. At the time, this system was more than adequate for the tasks at hand, but due to the increase of data that needs to be kept secure and away from prying eyes, this model is severely outdated.
As such, the job of a data centre architect has becoming increasingly difficult. Essentially, while the compute stage of data processing is now more important than ever, architects have to think about the movement, storage and processing of data first because the industry has become data-centric.
Recent developments have lowered the immediate issue, but caused further ones
Fortunately, the weight has been lifted off of the shoulders of the architecture of storage facilities somewhat due to recent breakthroughs in big data. Now, rather than compute-centric systems, rack level optimisation allows for data-centric systems.
In order to achieve rack data optimisation the following three points have to be carried out:
- High speed data (HSD)
- Effective data movement
- A separation of the servers resources to optimise performance
Rather unfortunately, the movement to data-centric architecture has caused issues to the networks that originally had little issue operating on the existing three-tier system.
Networks usually work at speeds from 1 gbps to 10 gbps, but this speed simply isn’t enough to complete even the simplest of tasks at peak times, which can cause long data transfer delays.
To counteract this issue, networking technologies have been improved with the introduction of technology that can provide speeds up to 100 gbps, which is 10 times the maximum speed previously offered.
However, this isn’t the only improvement. In fact, new networking technology actually allows for a direct transfer of data to the application without using up memory or CPU resources, which is arguably more important.
The architecture of Cloud computing
Cloud computing architecture is ‘made up’ of four components; front end, back end, Cloud based delivery and a network (usually the Internet), and how it is delivered can be broken down into a further four services; Software as a Service (SaaS), Development Platform as a Service (DPaaS), Platform as a Service (PaaS) and Infrastructure as a Service (Iaas).
The infrastructure of Cloud Computing
The infrastructure of Cloud Computing needs to follow four core principles in order to be successful, including; security, management, scalability and transparency.
Security is one of the priorities, especially as many people remain sceptical about how their documents are stored ‘in the Cloud’. However, what this actually comes down to, and to give a hint to the answer to the original question, is often the security of the ‘mega data centre’.
Not only should one be concerned about application security, which should be a priority, but also with how the servers are secured. In any instance, any data centre should have top of the line protection from those who may be looking to corrupt the data at its source.
What comes first?
So, taking all the above information into account, and looking at that last section, you should be able to tell what comes first, the architecture or infrastructure.
The answer is architecture. This is simply because the layout needs to be correct in order for the infrastructure to be enhanced to its fullest potential.