This article originally appeared on Inside-Cloud.com on November 28, 2012.
Ever since the timeshares of mainframes and the arrival of PCs – the power of massive computing has been in the hands of the few. However, the world is different now.
You may read about the “consumerization of IT” – where employees are bringing not only their own mobile devices, but they are bringing their own laptops and, in some cases, software to work. These employees don’t feel a sense of “entitlement,” they feel a sense of empowerment. You may also remember the teams that built the data warehouses of the 90’s and how they then leveraged the easy to use Business Intelligence tools of the first part of this century. Employees in many departments are now curious. Very curious, and no IT department or budget is going to get in their way.
They have a lot of email stored, they have a ton of music stored, they index their laptop of work files – and they write requirements to store all data – beyond what is required by law. Data is flowing within their organization at an ever increasing rate. They want to apply their “consumerization” of IT to the data at their fingertips.
What’s different now? Cloud computing? No, not first generation difficult-to-use, costly, slow and limiting – but second generation Infrastructure as a Service (IaaS). We now live in a world where instances can be connected at 80Gb/s, and instances can have 196GB of RAM and 48 cores. The real clouds are arriving. And despite what some might think, they are priced and packaged for the masses.
Competition in every industry is stronger and growing faster now too. It’s fueled by data. The early adopters of cloud computing had a competitive edge.
“Now, it’s time to democratize compute power.”
Data is all around us and should be available when and where we need it. Used efficiently and effectively, data can greatly improve how business is done. It improves product development and increases the bottom line.
Gmail. EDI Streams. Retail Transaction Data. Flickr. Amazon. WorkDay. SalesForce.com. Citibank. YouTube. Data is constantly moving into the cloud and being absorbed at a record rate.
How is this possible?
Instead of running on a physical machine, today’s modern applications work with data run on the cloud in giant compute and storage clusters of virtual machines, created by software and controlled by APIs.
At some point, cloud computing becomes physical. For anything to happen, a CPU will have to load instructions and process the data. To initiate the processing, the data has to arrive at the computer containing the CPU and will then be sent somewhere, often to another computer. This is where the physical network connecting the computers comes into play. It is in the switching layer that data flow challenges will appear as the computing world becomes “virtualized.”
However, for all of the great strides in virtualization and decreasing hardware costs, for many organizations High Performance Computing environments have not been an option. CPUs, Memory, Storage and Network interconnects were simply too costly to acquire and maintain.
The promise of cloud computing in its first generation did not improve the performance, cost or ease of use to the HPC market. The biggest stumbling block to date has been the cost in both performance and I/O for networks in the cloud.
As an IaaS provider, ProfitBricks has developed what we believe is an elegant solution to the “network traffic roadblock”. At the core of our solution you will find InfiniBand, a wired fabric that allows more than triple the data transfer rate between servers as compared to two of the industry’s biggest players, Amazon and Rackspace. ProfitBricks uses Mellanox dual QDR 40Gb/s InfiniBand switches to connect the server and storage infrastructure, while using a KVM hypervisor and massive multi-core AMD Opteron processors.
Industry Example – Leveraging Cloud Saves BioPharm Time and Money
One place where cloud computing is growing rapidly in importance is within life sciences, particularly R&D organizations that are deluged with data from a variety of sources. At the same time, demand for computationally complex modelling and simulation studies continues to increase dramatically. Limited funding and budgets make it difficult for many organizations to build the IT infrastructure necessary to keep pace with these challenges. For many companies, cloud computing appears as a promising alternative to in-house expansion of IT services.
The main impact to the increased usage of cloud computing by pharmaceutical companies is reduced dependency on their own in-house IT infrastructures. Offering the ability to implement services in a matter of minutes, by self-service from a Web platform, cloud computing also provides the ability for these companies to move away from intensive capital expenditures to an operating expenditure/pay-as-you-go business model. The business advantages of cloud computing include the standardization and streamlining of operations, and since all parties have access to the cloud, stronger collaboration among external entities and the healthcare ecosystem.
Cloud computing has clearly emerged as an alternative computing platform that bridges the gap between scientists’ growing computational demands and their computing capabilities. The cloud allows scientists, researchers, and other big data crunchers to obtain massive computing resources within minutes, compared to a number of days or even weeks it normally takes under traditional business processes.
A research study published by Open Research, and led by Qiming He, affiliated with the Illinois Institute of Technology, reveals:
- Virtualization technology, which is widely used by cloud computing, adds little performance overhead;
- Most current (first generation) public clouds are not designed for running scientific applications primarily due to their poor networking capabilities.
However, a cloud with moderately better network (vs. EC2) will deliver a significant performance improvement. Newer HPC-friendly “science cloud” platforms will be equipped with much better processing and networking capabilities such as InfiniBand, the platform behind the US Department of Energy’s Magellan and cloud computing IaaS providers like ProfitBricks.
HPC in the cloud is a proven advantage for the BioPharm industry, but it won’t stop there. This more cost-effective and efficient way to improve computing performance is available now. It was the vision of ProfitBricks’ founders to offer the second generation of cloud to every organization. And now that it’s here, there’s no stopping it.