Dell EMC officials are introducing fully integrated and pre-validated systems that come with the hardware and software enterprises need to more quickly begin running artificial intelligence workloads.
The company this week unveiled two new offerings in its Ready Solutions lineup that are specifically designed for machine learning and deep learning tasks and save businesses the cost and trouble of pulling together and integrating the various components themselves. The systems are Machine Learning with Hadoop and Deep Learning with Nvidia.
They’re the latest systems coming to market designed to make the adoption of AI and its subsets of machine and deep learning easier. They come complete with the technologies needed and are optimized to accelerate AI workloads.
Other system vendors are making similar efforts to reduce the complexity associated with AI by offering pre-integrated and pre-validated systems. Most recently, NetApp earlier this month partnered with Nvidia to develop the ONTAP AI architecture to help enterprises better manage the huge amounts of data that is being generated.
AI promises to reshape the how businesses are run, and most companies are looking to bring it into their environments. Analysts at the Enterprise Strategy Group (ESG) said that 69 percent of respondents to a recent study expect machine learning and AI to deliver measurable outcomes in the near term, and 17 percent said the technologies are critical to their companies’ strategy. Dell EMC officials said research done with market research firm Vanson Bourne found that 80 percent of organizations will be investing in AI technologies within the next five years.
At the Dell Technologies World show in April, CEO Michael Dell said he is seeing an “absolute explosion for use of AI. The idea here is you’re taking all this data and learning an inference to draw better conclusions from the data. We’ve seen a rapid acceleration in our server business last quarter and this quarter. When you dig into why this is happening, you’re seeing AI as a big use case.”
However, bringing the technology into a business environment can be complex and expensive, according to Jack Poller, senior analyst with ESG.
“Lacking a standardized AI infrastructure stack, organizations can invest the time, effort, and money to select, acquire, integrate, configure, test, and validate their own custom stack,” Poller wrote in a technical review of the Dell EMC systems. “This complex process can take months, and the organization must juggle purchasing and support across many vendors. Public cloud solutions suffer from huge cost variability and the time and cost necessary to transfer and store terabytes to petabytes of data.”
Both of Dell EMC’s offerings are based on the company’s PowerEdge servers and include not only compute, storage and networking capabilities but also libraries and AI frameworks like Caffe and TensorFlow.
The Deep Learning with Nvidia system uses PowerEdge 740xd and C4140 servers that include four Nvidia Tesla V100-SXM2 Tensor Core GPU accelerators, which are designed to drive deep learning performance. The V100, with 640 cores, was released last year and was designed for AI workloads, delivering up to 100 teraflops of performance. The Dell EMC system also includes the company’s Isilon F800 all-flash NAS storage and Bright Computing’s Cluster Manager for Data Science combined with Dell EMC’s Data Science Provisioning Portal.
The Machine Learning with Hadoop solution expands the vendor’s Ready Solutions for Hadoop, which were created with Cloudera and Intel. The machine learning offering includes Dell EMC’s R640 nd R740xd systems, the Cloudera Data Science Workbench, the Apache Spark data analytics engine and the Dell EMC Data Science Provisioning Engine, which brings with it preconfigured containers for data scientists to access Intel’s BigDL distributed deep learning library on the Spark framework.
The new offerings are designed to improve data science productivity by up to 30 percent and reduce the time needed to get an AI environment running by six months to a year compared with organizations that try to bring all the components together themselves, according to company officials. A data scientist’s workspace can be configured in five clicks, they said. The officials also said the Ready Solutions for AI will deliver twice the performance of competitive offerings.
The systems are available now in the United States and will launch in other countries, including Canada, Mexico, Brazil, France, Germany, the United Kingdom, China and India, within 60 days. The solutions also come with deployment services from Dell EMC.