GPU based servers can help solve Big Data energy woes
|Richard Harris in Big Data Tuesday, May 22, 2018|
GPU based servers are more energy efficient, and perform better than their CPU based models. We gained insight from Ami Gal, CEO of SQream, into Big Data storage problems data centers are facing today and how they can downsize their footprint but upsize their performance.
ADM: Why are data centers still growing?
Gal: As enterprise data grows exponentially, so too do the complexity, latency, and shortfalls of disseminating and ingesting the data for the cloud or for local data centers. Enterprises are working hard to keep the data closer to where it is generated and where it is consumed. So not only are data centers growing in many cases, but we are seeing the distribution of data in the cloud and across data centers where data is stored and analyzed. The real challenges begin when you need to merge this data and analyze data stored in multiple locations and data stores at the level of terabytes and petabytes.
ADM: Doesn't the move to the cloud help to reduce data centers overall?
Gal: While there has been a lot of talk about the move to the cloud, and in fact, organizations have already taken steps to move their applications and/or data to the cloud, many enterprises are not planning on moving to full cloud operations.
Enterprises are often creating hybrid cloud architectures, merging existing or new data centers with integration into the cloud. The phenomenon of organizations using the hybrid model to breach the cloud model as the first step to a longer term and fuller cloud implementation is still not clear. The hybrid model works well for many enterprises, especially for those wishing to keep certain data stores close to the data’s origin or where it will be analyzed.
ADM: Why should organizations care that their data centers are environmentally friendly?
Gal: While it would be nice to say that enterprises are altruistic when it comes to the environment, this may be true only for some organizations. Regardless of whether the driver comes from noble desires to help the world or not, organizations can benefit greatly from being environmentally friendly.
It is typical in enterprises that the responsibility for data center costs often falls in the hands of the facilities management department. IT management is often unaware of data center energy costs and other facility-oriented costs such as those related to real estate.
Taking a "green" outlook on data centers can bring significant benefits in reducing energy costs, real estate costs and associated facility management. But the benefits of going "green" extend beyond saving money or helping the planet.
By using advanced computing innovations, like GPU-based servers and databases, much less hardware is needed, making maintenance significantly easier and less time-consuming. In addition, the ability to respond to business needs and advanced analytics is much greater. So not only are IT investments reduced and overall costs decreased, but the business benefits overall.
ADM: What are the options available for companies who want to downsize their data center?
Gal: Enterprises who want to downsize their data centers have various alternatives at their disposal. They may consider moving some or all of their applications and data stores to the cloud. Even if they make this decision, it is not done in one shot, and often begins as a short-term hybrid model with a structured plan to move fully to the cloud in the long run, which may or may not happen.
Regardless of whether an organization decides to move to the cloud, a move to more innovative processing methods can help significantly downsize computing resources, whether within an existing data center or in the cloud. By embracing new GPU computing technology, and massive parallel processing, for example, enterprises are able to significantly reduce their number of servers, which in turn reduces costs associated with energy, space, maintenance and more.
ADM: Why are GPUs more environmentally friendly than CPUs?
Gal: GPUs increase energy efficiency, reduce the need for large air-conditioned data centers, and significantly decrease the cost of maintenance. The goal of every IT manager is to increase performance and service quality to business users while reducing costs.
Originally used in video gaming, GPUs have become the most energy-efficient processors in the market, significantly more power-efficient than CPUs. They use parallel processing to break down complex problems into smaller tasks that run simultaneously. This enables a single GPU, for instance, to do the work of multiple CPU servers, consuming much less energy and greatly reducing maintenance costs and resources.
Many enterprises are already adapting GPU servers and databases to realize dramatic gains in energy efficiency while increasing the performance of their applications and data analytics.
ADM: What's the future of big data?
Gal: Data is growing at unprecedented rates and will continue to grow exponentially. Enterprises will have to adopt new technologies that will enable them to more effectively and efficiently store and analyze growing data stores. Databases created twenty or more years ago were not designed for massive data analytics, and therefore aren’t capable of effectively ingesting and analyzing data stores of tens to hundreds of terabytes or petabytes.
In addition to the growing data challenge, demand for data scientists, analysts and data management experts will spike. Expenditure on managing and analyzing big data will grow to billions of dollars annually as industries such as banking, retail, healthcare, and internet-based businesses increase their big data spend, mostly on technology to help manage and leverage their data.
Having these very large data stores will not provide enterprises with competitive advantage unless they are able to effectively and rapidly analyze the data and extract actionable intelligence from it. To facilitate this requirement, companies will continue to adopt databases that can enable them to rapidly analyze their data. With GPU-based databases, for example, organizations can quickly analyze massive amounts of data to gain a competitive advantage.