Scaling new HPC with Composable Architecture

[Disclosure: I was invited by Dell Technologies as a delegate to their Dell Technologies World 2019 Conference from Apr 29-May 1, 2019 in the Las Vegas USA. Tech Field Day Extra was an included activity as part of the Dell Technologies World. My expenses, travel, accommodation and conference fees were covered by Dell Technologies, the organizer and I was not obligated to blog or promote their technologies presented at this event. The content of this blog is of my own opinions and views]

Deep Learning, Neural Networks, Machine Learning and subsequently Artificial Intelligence (AI) are the new generation of applications and workloads to the commercial HPC systems. Different from the traditional, more scientific and engineering HPC workloads, I have written about the new dawn of supercomputing and the attractive posture of commercial HPC.

Don’t be idle

From the business perspective, the investment of HPC systems is high most of the time, and justifying it to the executives and the investors is not easy. Therefore, it is critical to keep feeding the HPC systems and significantly minimize the idle times for compute, GPUs, network and storage.

However, almost all HPC systems today are inflexible. Once assigned to a project, the resources pretty much stay with the project, even when the workload processing of the project is idle and waiting. Of course, we have to bear in mind that not all resources are fully abstracted, virtualized and software-defined whereby you can carve out pieces of the hardware and deliver a percentage of that resource. Case in point is the CPU, where you cannot assign certain clock cycles of CPU to one project and another half to the other. The technology isn’t there yet. Certain resources like GPU is going down the path of Virtual GPU, and into the realm of resource disaggregation. Eventually, all resources of the HPC systems – CPU, memory, FPGA, GPU, PCIe channels, NVMe paths, IOPS, bandwidth, burst buffers etc – should be disaggregated and pooled for disparate applications and workloads based on demands of usage, time and performance.

Hence we are beginning to see the disaggregated HPC systems resources composed and built up the meet the diverse mix and needs of HPC applications and workloads. This is even more acute when a AI project might grow cold, but the training of AL/ML/DL workloads continues to stay hot

Liqid the early leader in Composable Architecture

Continue reading

Disaggregation or hyperconvergence?

[Preamble: I have been invited by  GestaltIT as a delegate to their TechFieldDay from Oct 17-19, 2018 in the Silicon Valley USA. My expenses, travel and accommodation are covered by GestaltIT, the organizer and I was not obligated to blog or promote their technologies presented at this event. The content of this blog is of my own opinions and views]

There is an argument about NetApp‘s HCI (hyperconverged infrastructure). It is not really a hyperconverged product at all, according to one school of thought. Maybe NetApp is just riding on the hyperconvergence marketing coat tails, and just wanted to be associated to the HCI hot streak. In the same spectrum of argument, Datrium decided to call their technology open convergence, clearly trying not to be related to hyperconvergence.

Hyperconvergence has been enjoying a period of renaissance for a few years now. Leaders like Nutanix, VMware vSAN, Cisco Hyperflex and HPE Simplivity have been dominating the scene, and touting great IT benefits and eliminating IT efficiencies. But in these technologies, performance and capacity are tightly intertwined. That means that in each of the individual hyperconverged nodes, typically starting with a trio of nodes, the processing power and the storage capacity comes together. You have to accept both resources as a node. If you want more processing power, you get the additional storage capacity that comes with that node. If you want more storage capacity, you get more processing power whether you like it or not. This means, you get underutilized resources over time, and definitely not rightsized for the job.

And here in Malaysia, we have seen vendors throw in hyperconverged infrastructure solutions for every single requirement. That was why I wrote a piece about some zealots of hyperconverged solutions 3+ years ago. When you think you have a magical hammer, every problem is a nail. 😉

In my radar, NetApp and Datrium are the only 2 vendors that offer separate nodes for compute processing and storage capacity and still fall within the hyperconverged space. This approach obviously benefits the IT planners and the IT architects, and the customers too because they get what they want for their business. However, the disaggregation of compute processing and storage leads to the argument of whether these 2 companies belong to the hyperconverged infrastructure category.

Continue reading