We are using AI (or Artificial intelligence) already all the time throughout our day, without being aware it. AI is present in our everyday life, every time we use SIRI, google assistant or Cortana, they know every aspect of our life – the coffee meetup with a friend every Sunday at the same time and place, our everyday commute details, our spouses’ birthdays or even the child’s next football match. Facebook’s software is even capable of recognizing famous faces and tagging them automatically without any human assistance.
According to a technology review by Google, in 2017 more than 26% of the interviewed companies planned to invest more than 15 % of their IT budget on machine learning and AI. AI is applied to solve bigger problems with a business purpose – for instance, a fast food chain is using AI to increase customer engagement and increase shop visit by pushing a targeted ad for a free ice cream on a hot afternoon. Another example of solving problems with AI – finance institutions are analyzing a large amount of data in order to discover frauds and eliminate risk. Machine learning on the other side plays an also important part at Uber. Uber uses machine learning algorithms to predict more accurate arrival times, pick up locations and so on. Furthermore, did you know that airplanes are already using AI for autonomous flying? Autopilot mode for instance is based on flight planning management systems, that defines the most efficient route from point A to be. The dedicated system analyses data from a large number of sensors and adjusts the plane speed, rate of climb, altitude accordingly throughout the flight.
As announced last week, NetApp and Nvidia have joined forces to bring to market a solution called NetApp ONTAP AI to enable companies to adopt artificial intelligence faster and easier. NetApp with its Verified Architecture Program offers customers a proven architecture which is thoroughly tested, easy and quick to deploy and within a short time to market. NetApp ONTAP AI combines NVIDIA DGX-1 servers with NVIDIA Tesla V100 graphic processing units and a NetApp AFF A800 (also lower models are supported) with ultra-fast networking with Cisco Nexus 3232C 100Gb Ethernet switches, that are useful for inter-GPU communications by remote access (RDMA) over Converged Ethernet (RoCE). Traditional HPC infrastructures are connected to compute nodes via RDMA over InfiniBand which is providing high-bandwidth and low-latency, but now when an Ethernet technology has the same or better performance. That’s the reason why NetApp ONTAP AI uses well understood and widely deployed Ethernet technology. Data on storage is accessed through NFS protocol using four 100GbE links. Each DGX-1 server is powered by eight Tesla V100 GPUs configured in a hybrid cube mash topology leveraging the NVIDIA NVLink technology. The NVlink technology provides low-latency fabric for inter-GPU communications. DGX-1 is powered by NVIDIA GPU cloud, providing Nvidia GPU optimized containers for the most popular DL frameworks like Caffe2, TensorFlow, PyThorch, MXNet and Tensor RT and is incorporated with NVIDIA CUDA toolkit. NVIDIA CUDA toolkit is providing NVIDIA CUDA Basic Linear Algebra Subroutines Library (cuBLAS) and the NVIDIA CUDA Deep Neural Network Library (cuDNN). NetApp ONTAP AI simplifies deployment by eliminating design complexity and guesswork. NetApp ONTAP IA solution is scalable up to multiple racks and is capable of providing a massive throughput up to 300 GB/s and 11.4 million IOPS in a 24-node cluster. Imagine how much faster and how much more data can be analyzed with a solution with this kinda performance. When your complete system depends on a solution, the solution must be capable to be upgraded non-disruptively. One of greatest and long living NetApp’s principles is that you can start small and grow with without interruption. Similar to FlexPod, the first solution collaboration between Cisco and NetApp, NetApp ONTAP AI comes along with a single point of contact support, backed by AI experts, a winning strategy that has been proven invaluable with FlexPod customers in the past.
I believe easy access and deployment of AI solutions will make huge changes in our lifes that we are currently not capable to understand and imagine. Hopefully in the future, I solutions like NetApp ONTAP AI will contribute to discovering cure for deadly diseases like cancer and HIV, growing food more efficiently, making transportation faster and safer and maybe even automating basic tasks like calling car service or making a dentist appointment, so we, humans, would have more time to spend with our loved ones.