People close to me know that one of my biggest passions is hiking. I believe that has a lot to do with my father taking us on multiple hiking trips, and it probably even goes deeper than that through our family roots. Ever since, I’ve been hiking across Israel as well as multiple continents and countries, including my honeymoon which was spent mostly in Chile and Argentina, hiking the amazing Andes.
The Big Bend
Last month, I had the opportunity to visit one of our prospects in Dallas, Texas for a technical session discussing our recently announced K2 Gen6 All-Flash storage array and the new software features of VisionOS 6.0. As it turns out, a close friend who now lives in Houston heard about my trip and insisted (it didn’t take much insistence, really!) that we should spend a couple of days hiking the Big Bend National Park. Knowing him well, I knew better than to question his plans and agreed. So, other than enriching your knowledge about the Chissos Basin and the South Rim (highly recommended for a hike, by the way!), you are probably wondering why I am talking about this! In a nutshell, it’s about scale.
The Big Bend is part of the Chihuahuan desert. Coming from Israel, I know that type of topology well. What I wasn’t prepared for was the sheer size and variety of features in that desert, much like our existing customers meeting our 6th generation platform and software for the first time.
Scaling new heights with K2 Gen6
For those who aren’t familiar with our all-flash storage product, the K2 is a fully Active-Active (not just frontend or for specific functions), scale-out and scale-up all-flash storage array. The basic building block of the K2 is a K-Block that is comprised of two controllers, SAS connected (12 Gb/s) to several SSD shelves. The flexibility of scale-up is in the idea that you can connect several shelves to add more capacity. The 6th generation of K2 now includes support for up to four shelves full of 4TB drives! That sums up to an impressive 384TB of RAW capacity. With our super-efficient K-RAID and data reduction technologies that map to GUARANTEED effective capacity of 960TB in just 10 Rack Units!
While some of our competitors may offer that capacity scale, what differentiates us is our capability to also scale out (which is super hard to achieve engineering-wise!). Scaling out indicates that we we can also scale such a system with more K-Blocks, and with our current tested and certified configuration of four K-Blocks, we can get to 1.5PB of RAW SSD, and 4PB effective capacity in a single fully Active-Active system occupying 42 Rack Units (IB switches occupy 2 RUs). That’s a high number. Much like the Big Bend Emory Peak’s 2,385m altitude compared to Mount Ramon in the Israeli Negev towering 1,037m.
In storage, people also like to quote performance numbers. In Kaminario, performance has always been extremely important, so this comes naturally. As such, the new platform boosts performance by using newer Intel processors, faster RAM and more. We’ve also upgraded our FC connectivity to 16Gb/s (32Gb/s upgradable) to now achieve an impressive 6.2 GB/s throughput in a single K-Block!
Hardware is great, but that’s basically just the chassis for our core competency, which is software!
Software driven architecture
Our great engineering team has been working not only on utilizing this new hardware (managing that amount of data and IOPS requires sophisticated data structures and software) but also on algorithm optimizations, data reduction techniques and a great deal more to continue to push the envelope on performance, flexibility and cost-optimization. This resembles the way my friend “optimized” our short time in the Big Bend, by pushing me not only to complete the ‘South Rim’ trek, but also to follow immediately with the ‘Lost Mine’ and top it off with a visit the Big Bend hot springs.
With V6.0, we also added the ability to automatically schedule snapshot creation and retention out of the box. K2 snapshots are super fast, take no extra space and do not affect performance. So now, instead of scripting and using cron or Windows scheduler + calling our REST APIs or CLI, one can simply define a “snapshot scheduler” in the K2 GUI, assign it to a volume group (or a bunch of them) and let it run. It will create the snapshots in the configured time, and of course perform the housekeeping of deleting those per the defined retention policy for this scheduler.
The next big challenge
The product and engineering teams at Kaminario come to work every day with a singular goal – make everything as simple as it can be. The K2 was built with the core principles of scale, performance and resiliency in mind but what underscores all of this is simplicity. Our challenge is to continue to refine and simplify our platform each and every day.
As with any great hike I’ve been part of, as soon as I’m done with it, I start thinking about my next challenge. At Kaminario, we’ve embraced that mission as well. While we may have just released a ton of new functionality with the Gen6 K2, we are already mapping out some incredible new exciting announcements in the coming months – stay tuned!