DaaS

Scalability: What Clouds Are Made Of

We have mentioned the feature “scalability” multiple times on the IronOrbit blog. We even briefly defined the term in our “Cloudy to Clear” series. In today’s post, we will focus in depth on scalability and its uses, benefits, and latest developments.

In the third edition of our “Cloudy to Clear” series, we wrote: “A system is scalable that can be ‘scaled’ up or down, or enlarged or shrunk, and still maintain its basic composition and proportions. A scalable cloud can add further users, storage, or processing power simply by connecting extra hardware to the existing infrastructure. By definition this type of expansion does not require any extended integration or reworking of the existing system and will not disrupt the service of ongoing users.” A scalable system has several benefits. Most importantly, a scalable solution retains its singularity and centralized management as it expands and contracts. While an un-scalable system can match the total performance or storage capacity of a scalable alternative, it lacks efficiency and the ability to perform wide-ranging tasks or impose consistent settings and policies. The singularity of scalable systems allow them to 1) perform specific, complicated tasks while consulting massive amounts of data; 2) ensure system-wide performance and security; 3) prioritize and maximize computing resources, letting richer companies get more out of an existing IT hardware infrastructure and poorer companies be able to afford a high-performance, high-capacity system.

Because of their combination of cost-effectiveness, high performance, and large storage capacity, scalable solutions form the foundation of technologies as various as clouds, virtual desktop infrastructures, research supercomputers, and web services like YouTube and Twitter. Scalability also has a lot to do with Big Data management and analytics. The term “Big Data” refers to extremely large amounts of related data and all the technical and ethical issues associated with the management, storage, and sharing of that quantity of information. For example, the social media website Facebook operates and manages about 10 full-size data centers that allow it to host 30 billion new pieces of content every month and 100 million new photos every day. Privacy advocates question both whether Facebook has the technical ability to securely manage their Big Data and also if a single non-government entity should possess and control so much information about so many people. Regardless, without a scalable IT infrastructure, Facebook would not have the same performance speeds or storage capacity and would probably have to charge for its services.

Reinforcing the utility and importance of scalability-backed Big Data technology, the Obama administration recently announced over $200 million in funding for Big Data research and development. Twenty-five million dollars of the total has been allocated for the Scalable Data Management, Analysis, and Visualization (SDAV) Institute, a partnership between a private-sector developer of Big Data analysis tools and the illustrious government research organization Lawrence Berkeley National Laboratory to improve the speed and relevance of the filtering of large scientific data sets. Scalability-enabled Big Data has already had a sizeable influence on the world of scientific research: the CERN particle accelerator produces almost 20 gigabytes per minute and 13 petabytes of data (about 13,500,000 gigabytes) every year, processing and storing the data with a cloud-based infrastructure with a total capacity of 150 petabytes and 200,000 processing cores.

IronOrbit utilizes in our own hosted infrastructures and applications the same kind of high-performance scalable technologies that sustain Facebook and the CERN particle accelerator. Thanks to scalability we can build low-cost, high-capacity solutions for our clients and add users and storage on demand. With scalable infrastructures and applications from the IronOrbit, your company’s growth will be supported and reinforced, not impeded, by the performance and capacity of your IT system.