Persistent, Portable and Performant Containers and K8s

According to Gartner Group, by 2022, more than 75% of global organizations will be running containerized applications in production.[1]  To maximize the potential for containerized applications, under any orchestration tool (Kubernetes, Docker Engine, etc.) the underlying storage systems must address three fundamental requirements:   Persistence Portability Performance   Persistence Containers were traditionally designed as ephemeral resources for development and test.  As container benefits like the separation of infrastructure from application becomes obvious, there is a natural desire to move from experiment to production.  This requires a definition of a specific container as persistent.  Using functions inherent to the Pavilion system,…

Maximizing NVMe Performance and Operational Efficiency

In their latest installment on NVMe storage systems, 451 Research spells out a clear vision for NVMe storage where NVMe-Over-Fabrics (NVMe-oF) is the clear winner for Web-Scale, Media and Entertainment, Governments and Financial Services firms with a need for speed.  Research Analyst, Tim Stammers identifies four pillars of NVMe deployment that merit further investigation:   Modifying existing storage systems to use NVMe does not maximize performance Offloading processing onto host servers drives up costs by preventing disaggregation Relying on applications for data services also drives up complexity and management overheads Architecting storage to maximize NVMe performance while retaining services maintains…

Ready for IBM Storage

Get ready!  Pavilion Data’s Hyper-Parallel Flash Array is now “Ready for IBM Storage.”  The first Spectrum Scale solution for shared NVMe storage is now formally listed on the IBM Global Solutions Directory.   With 120GB/sec read and 90GB/sec write bandwidth at 40 microseconds, Pavilion Data offers performance density unparalleled in shared storage for Spectrum Scale, and we have several significant deployments.   A Massively Parallel Computing Center has deployed our storage array as part of a genomics sequencing cluster running analytics through more than 120 nodes with a single 4U chassis.   Another customer has more than 512 nodes deployed…

Top 8 Reasons to choose Pavilion Data’s Hyper-Parallel Flash Array

With listels all the rage these days, I thought it would be interesting to summarize the key points from 2019’s previous blogs and cover a few more in this BuzzFeed friendly format.  While it’s not exactly my style to boil deep technical insight into a top 8 listel, it is good to be concise.   Customers are voting with their pocketbooks. Since the beginning of 2019, Pavilion Data has shipped dozens of systems to customers around the globe and closed multiple 7-figure deals.  Most deals several things in common.  First, a never-ending thirst for speed that only NVMe can quench. …

Hollywood Post-production & Pavilion Data

Full uncompressed editing with multiple collaborators requires massive storage bandwidth and low-latency.  Traditionally, this is achieved by purchasing a large number of storage arrays and disks (HDDs or SSDs).  The result is excess capacity stranded for a single media asset.  This can be as much as four times the amount of actual storage required.  Pavilion Data can deliver 120GB/sec of throughput, enabling up to 10 seats with 6 streams of 60fps uncompressed 4K against an exact size of a media asset.   Most vendors are now promoting “All-NVMe” as the answer for uncompressed 4K and 8K editing and delivery.  But…