A key trend in the ever-changing IT world is bringing everything that is needed (processing, storage, analysis, networking) closer to where the data resides. The increasing amounts of data being generated and the number of datasets created – and the valuable business information captured in all of it – is putting pressure on compute, storage, and network vendors to find ways to reduce the time needed and the costs involved in processing and analyzing this data.
“This is an evolution underway in the industry, from a traditional system very focused on computing, where everything goes through the processor and all the computing tasks are managed there in this central CPU chipset and DRAM build ”JB Baker, vice president of marketing at computer storage vendor ScaleFlux, recounts The next platform. “Over time, what we’ve seen is that a few things are happening. First, domain-specific computing is able to handle certain tasks much more efficiently than processors. For example, GPUs and TPUs for AI workload acceleration, and now SmartNICs have come out to offload some network processing which can be much better handled in domain-specific silicon than in a CPU.
Computer storage “is the next wave of processing tasks that are not handled well in a processor, such as compressing, which can be extremely heavy if you are trying to run a high compression level algorithm such as a file format. gzip, or things that are going to have to happen on most of the mass of data which by doing these tasks in the drive, such as filtering a database, we can reduce the amount of data that must have happened. move on the PCIe bus and pollute the DRAM or even across the network, as networks are another bottleneck, “says Baker.” It’s really this evolution of domain-specific computing and the only mass of data and the speed of data that creates all of these bottlenecks in your compute infrastructure. ”
As we wrote, computer storage – also known as “in-place processing” – is an emerging space led by startups in a segment of storage that has solved capacity and speed issues over the years. the last decade, but is still challenged by cost and time. involved in moving it to where it can be treated. The idea is basically to process the data close to where it is already stored.
ScaleFlux is one such startup, launched in 2014 and since then raising $ 58.6 million in six rounds of funding, the most recent being last year, when the company grossed over $ 21.5 million. , according to Crunchbase. Customers include Nvidia, Tencent and Alibaba and technology partners include Hewlett Packard Enterprise, Dell EMC, Lenovo, Inspur and Super Micro.
The publisher launched two generations of products, the first in 2018 and the second two years later. ScaleFlux Computer Storage Disks (CSDs), which bring together PCIe SSDs and integrated compute engines, which offload data-intensive tasks from compute to free up processors and reduce I / O bottlenecks that arise. produce when sending huge amounts of data from storage devices to data center systems or disaggregated compute nodes.
Compute engines leverage field programmable gate arrays (FPGAs) and drives can be installed in standard slots for 2.5-inch U.2 drives or half-height, half-length PCIe expansion cards . IT administrators can also change settings to better balance storage space, performance, latency, and application endurance.
“Our main goal is to use drives that are better than your regular NVMe SSD,” says Baker. “We’ve embraced the medical mantra of no harm: make sure we deliver the performance, latency, endurance, reliability, all the features you expect from your enterprise NVMe SSD, then add innovation. to give you further improvements to your system efficiency and your training efficiency.
Now the ScaleFlux is coming out this week with its third generation of products, which are now in beta and will be available in spring 2022. The latest generation includes not only improved computer storage systems, but also Arm-based silicon and a sequel. discrete software designed to facilitate the use of computer storage by businesses.
“CSware is a software and application integration that understands the transparent compression capacity of drives and leverages it to improve performance, reduce CPU load, and reduce DRAM pollution,” says Baker.
There is software RAID in the CSware offering that will take advantage of disk compression to deliver RAID 5-like storage costs and RAID 10-like performance, he says. There’s a compression-optimized KV store that Baker says is an alternative to using RocksDB as a storage engine for a database, halving the amount of CPU and RAM consumed and doubling the performance of write the database.
A third piece of software can be used in relational databases such as MySQL and MariaDB to manage data updates, improve write performance to these databases, and reduce disk wear.
ScaleFlux will likely offer a free version of the software and donate it to the open source community to help drive adoption of computer storage.
The company’s new SFX 3000 Series storage processor and the firmware it contains are the computing power behind the vendor’s latest systems. The CSD 3000 series NVMe reader is the latest generation of the company’s flagship offering. It enables businesses to reduce data storage costs up to three times and double application performance. It also increases the endurance of flash memory by nine compared to other drives on the market, says Baker.
With four programmable Arm cores, enterprises can deploy application-specific distributed compute functions in the storage device.
“In the CSD 2000 generation that ships today, we have the capacity multiplier capability,” he says. “We have excellent performance in storage. We don’t have the encryption and we’re not using the NVMe driver yet. We have a ScaleFlux driver. In the 3000 [generation], we are moving to NVMe drivers to very simply support a wider range of operating systems and expand the market for these drives much more widely.
The NSD 3000 Series NVMe SSDs are essentially a smarter offering than other devices, with twice the endurance and random write and mixed read / write performance of competitive NVMe drives. It is aimed at businesses that may not need to embrace all of the capabilities of the CSD 3000 and are looking for a cheaper alternative.
“For those who aren’t sure exactly how they would consume the extended capacity or how they would consume those cores, here is a product where we disable those things and it’s just a better NVMe SSD,” he says. “In the NSD, we keep the transparent compression-decompression function, but we just use it to improve endurance and performance and not give customers the keys to multiply capacity. “
The CSP 3000 provides a built-in flashless computer storage processor and gives organizations extended access to programmable cores so they can add additional features specific to their deployments. This offer can be used when companies “have hard drives and they’re not looking to swap hard drives wholesale, but they want to perform a compute offload to go with that array or some other reason, ”says Baker. “If they need to separate compute from storage, then CSP is a viable solution. “
The development of the Arm-based ASIC was essential in enabling ScaleFlux to expand the capabilities of its portfolio, which could not be done with an FPGA in terms of cost and power, he says. The system on a chip (SoC) has four integrated functional zones. There is a flash subsystem with a 16-channel controller and built-in compression / decompression and encryption. Hardware compute engines are also integrated for fixed algorithm tasks and a CPU complex for use case specific programs. In the CSD, users have access to four of the eight cores; the others are used for flash management and exception handling. However, in the CSP, they have access to all eight cores before any flash is involved.
There is also a PCIe interface, with eight PCIe 4.0 ports, although in most form factors there will only be four lanes used.
“Everything is integrated into a single chip,” he says. “We don’t need a separate processor to do the compute functions and a discrete flash controller. Everything is integrated to save card space, improve performance, save power [and] allow greater disk capacity.
This data-driven architecture offers a range of advantages over traditional processor-driven compute infrastructure, including the ability to move more of these compute functions into disks, says Baker. What ScaleFlux seeks to do is provide businesses with a way to integrate even more of these computational tasks into their hardware.