Pavilion Data - IT Press Tour Jan 2021

Pavilion Data Systems. ©2021 Proprietary & Conﬁdential. All rights reserved.
IT Press Tour January 29, 2021

IT Press Tour Agenda Timing & Topic Presenter Take-Away(s) 7:00-7:02 Kick off Amy Love" Chief Marketing Officer Approach for the time 7:03-7:13 Welcome & Vision Gurpreet Singh Chief Executive Officer Achieving success and accelerating 7:13-7:20 Technology Vision V.R. Satish" Chief Technology Officer & Co-Founder Unrivaled Product Leadership that is Faster than the Speed of Light 7:20-7:33 Faster Than The Speed Of Light Costa Hasapopoulos Chief Field Technology Officer HyperOS 3.0 and unlimited scale and flexibility (Product Tech & IP) 7:33-7:43" Universally Unmatched Storage (product demo) Sundar Kanthadai VP Engineering & Co-Founder Customer engagement with the product 7:43-7:55 Q&A Open Q&A Ask away Timing & Topic Presenter Take-Away(s) 7:55-8:05 Unrivaled Choice & Control Amy Love" Chief Marketing Officer Positioning: " Market leadership, Customer Choice & Control 8:05-8:25 Universally Unmatched Solutions Costa Hasapopoulos VP WW Alliances Solutions being build & Partners(NVIDIA, OmniSci, Sony) 8:25-8:35 GTM/Market Success Chris McBride Chief Revenue Officer Unmatched Clear separation in Analytics, VMware, M&E (LB/LF) 8:35-8:45 Q&A Open Q&A Ask away 8:50-9:00 Conclusion Gurpreet Singh Chief Executive Officer On a mission!

Welcome & Vision Gurpreet Singh - Chief Executive Oﬃcer Achieving success and accelerating

Pavilion by the numbers 4 2018 Product GA 2x 2019 – 20 Revenue Growth $500m+ TAM across 10 customers $30B AI/ML/HPC/Analytics/Edge/IOT 30 Patents 3 Global Oﬃce Locations

Seasoned Leadership Team David Black VP Finance Gurpreet Singh CEO VR Satish CTO, Co-Founder Sundar Kanthadai VP Engineering, Co- Founder Chris McBride CRO Costa Hasapopoulos CFTO, WW VP Biz Dev Amy Love CMO Farzad Haghighi VP Customer Success Mike Gustafson Chairman, BoD Cloud Network Components Enterprise Storage $20B+ Market Value Creation 5 Ajay Patel VP Bus Operations

Business insights and decision making relies on modern applications 6 Modern Applications are built on a new, advanced, parallel application toolchain stack Applications Market Leaders Characteristics •  Small Data Volume: 100TB •  Monolithic, batch/transactional – Client/Server Architecture •  100x More Data: 10s of PBs •  High performance, real-time, Parallel and Scale outs architecture ERP HCM SCM Relational Databases Risk and Fraud Image Recognition Big Data Analytics 3D Volumetric Capture Genomics Traditional Applications Modern Data Intensive Applications

Massive Advancements in Compute, Memory and Network Technology to Enable Modern Applications Data IO Performance has Become the New Bottleneck 7 Improvement over Last 10 Years Server CPU Memory Storage Network Speed Throughput Latency (800x) (100x) (40x) (10x) (2x) Improvement Over the Last Decade

Far Larger Opportunity Than a Traditional Storage Focused Company 8 $18,0 $12,0 $35,0 $65,0 AI/HPC Infrastructure Spend Edge Computing Traditional Storage Total TAM ($bn) 2% CAGR 10% CAGR 24% CAGR ~8% CAGR Near-Term Growth Opportunity Future High- Growth Opportunity What is high performance now will quickly become mainstream Disrupt Large Existing Market

Unprecedented Scale multiple PBs of data under single global namespace Blazing Performance 10x-80x faster compared to “best-in-class” Unmatched Flexibility Multi-protocol and Future Proof Class-leading Density Compact space and power footprint The Pavilion HyperParallel Data Platform T A s Ha l HyperParallel Data Platform 9

Proven Success Across a Range of End Markets Finance & eCommerce Intelligence Government Life Science Media & Entertainment SaaS/ Internet 10

Technology Vision V. R. Satish - CTO and Co-Founder Unrivaled Product Leadership That Is Faster Than The Speed Of Light

12 Pavilion’s Rendering of the Future… Code AI/ML Compute Field Programmable Data Appliance Storage AI/ML Accelerators Computational Storage Controllers

13 Hyperparallel Platform Ultra Low Latency, High BW Network Distributed Shared Memory Devices Foundational Patent, the only low latency hyperparallel platform ... ... A Network Centric Design: N x N Connectivity

14 Hyperparallel Storage Array Ultra Low Latency, High BW Network SC SC Distributed Shared Persistent Memory Ethernet/IB A Network Centric Design: N x N Connectivity SC SC SC

15 Towards Pavilion’s True North Data Management Offloading Hyperparallel Storage Array Compute Offloading Hyperparallel Computational Storage Array Inference Offloading Hyperparallel Programmable Data Appliance Information

16 Programmable Controller Pavilion Extensions (PEx) Containers Network SRIOV HyperOS  Storage Services Control Plane Services HyperOS  version 3.x PEx Application Services Redundancy & Resiliency High Availability Distributed Persistent Memory Core Data Services Distributed Control Plane Quorum Services Metrics Alerting Orchestration Configuration Services Web UI CLI Pavilion Extensions SC

17 Pavilion Extension Interface: Seamless Integration Navigation Node Inserted by PEx UI Dashboard Portlet Inserted by PEx UI Details Page Inserted by PEx UI

18 Pavilion HyperOS 3.0 Blocks High Capacity – Scale-out – Low Latency – High Bandwidth Standards Based Hyperparallel Files Objects

SC 19 Containers Network SRIOV PavilionOS  Storage Services Control Plane Services PavilionOS  version 3x PEx Hyperparallel Filesystem Redundancy & Resiliency High Availability Distributed Persistent Memory Core Data Services Distributed Control Plane Quorum Services Metrics Alerting Orchestration Configuration Services Web UI CLI Pavilion Extensions NFS Programmable Controllers Hyperparallel Filesystem

20 HyperParallel Programmable Data Platform Storage Computational Storage Controller Personalization using PEx AI/ML Accelerators

Faster Than The Speed Of Light Costa Hasapopoulos - Chief Field Technology Oﬃcer & Worldwide VP Business Development HyperOS 3.0 and unlimited scale and flexibility (Product Tech & IP)

Why Are We Diﬀerent? 4 Rack Units 4U 6.9” 20 Independent Storage Controllers (CPU, Memory & Network) 40 x 100 GbE/HDR or 10 x 200 EDR/HDR Unique Architecture: •  Built like a network switch – PCIe Backplane •  Built from the ground up to support NVMe and NVMe-oF •  Multi/Many Controllers 72 x 2.5” U.2 NVMe Drives PCIe Switch based architecture (6.4Tb) •  Any controller to Any Drive Connectivity •  Cache-less and Tier-less •  DMA from any drive and any controller (nanoseconds) •  RDMA from HOSTS to Pavilion (microseconds) •  Software patents that take “unfair advantage of our hardware”

The RESULTS – The Most Flexible High Performance Storage Platform File Object Block DENSITY Multi-Storage Type / Protocol Best in Every Class Performance NFS v3 NFS v4 NFS RDMA NVMe- RoCE NVMe-TCP NVMe-RDMA iSCSI S3 90 GB/s – Read 56 GB/s - Write Read 120 GB/s 20M IOPS @100 µsec Write 90 GB/s 5M IOPS @ 25 µsec 80 GB/s – Read 35 GB/s - Write Not just Performance Density But Flexible Performance Density All at market competitive pricing 4 Rack Units 2.2 PB RAW 3200 Watts 20 Million IOPS 20 Controllers 40 100 Gb Ports 540 TB per RU 5M IOPS per RU 800 Watts per RU 5 Controllers per RU 10 x 100Gb per RU File Object Block

Deployment Options – The Most Flexible High Performance Storage Platform Block Solutions Not just Performance Density But Flexible Performance Density All at market competitive pricing Block … … Block Object … … File External File Systems via Block High Performance Distributed Object Solutions High Performance Distributied File Solutions Priorities for Pavilion HyperOSTM 3.0

Pavilion HyperOS version 3.0 – Enhancements to Block, File and Object ü  Container Support ü  Enhanced QoS support •  Volume Level •  Cap based QoS •  4 Tiers including “ludicrous mode” ü  2.2PB Raw ü  High Availability Enhancements • N+X Controller Availability • Data Assurance Enhancements ü  Windows NVMe-oF Driver Block File Object

Version 3.0 Pavilion OS – A Foundational Upgrade with Focus on File & Object NVMe-RoCE NVMe-TCP NVMe-IB iSCSI File System NFS v3 S3 NFS v4 Pavilion HyperOS  Storage Services Control Plane Services Pavilion HyperOS  version 2x Block, File, Object Global Namespace/ NFS Global Namespace/  S3 NVMe-RoCE NVMe-TCP NVMe-IB iSCSI Pavilion HyperOS  Storage Services Control Plane Services Pavilion HyperOS  version 3x NEW Block, File, Object Containers ü  Container Support ü  Multi-HFA FILE & OBJECT Chassis Namespace / Bucket

Pavilion HyperOS version 3.0 – Enhancement to Block, File and Object ü  Enhanced QoS support •  Volume Level •  Cap based QoS •  4 Tiers including “ludicrous mode” ü  2.2PB Raw ü  High Availability Enhancements • N+X Controller Availability • Data Assurance Enhancements ü  Windows NVMe-oF Driver Block File ü  Multi-controller / multi-chassis distributed ﬁle system ü  Single NODE Performance (scales linearly per chassis) •  75+ GB/s Read •  50+ GB/s Write ü  Single or Multiple Namespaces ü  Flexible Client Connectivity •  NFS V3/V4 •  NFS RDMA •  Gluster Native Client •  pNFS •  Hadoop & Spark plug-ins ü  Tiering ü  Replication ü  Snapshot & Clones ü  Security ü  Encryption Object ü  Multi-controller / multi-chassis distributed ﬁle system ü  Single NODE Performance (scales linearly per chassis) •  60+ GB/s Read •  40+ GB/s Write ü  Cloud Native Object Storage ü  Single or Multiple Namespaces ü  File Access Available ü  Worm ü  Tiering ü  Replication ü  Snapshots & Clones ü  Application Integrations / plug-ins ü  Replication ü  Security ü  Encryption

Flexibility of Deployment with Best in Class Performance Density Bring YOUR OWN file system Use OUR file system Use OUR block Use OUR object NFSv3 NFSv4 NFS RDMA Gluster Client pNFS Plug-in Plug-in SMB* File S3 Plug-ins Object NVMe-TCP NVMe-RoCE NVMe-RDMA iSCSI Block

Version 3.0 - High Performance File Systems (NFS and S3) NFS NFS S3 S3 NFS NFS NFS S3 NFS S3 S3 … … NFS … S3 NFS Single Pavilion Chassis Multiple Pavilion Chassis Cluster

Multi-Client OPTIONS and Interoperability – Ultimate Flexibility … Pavilion HyperParallel File System Single Mount Point //PavilionPHFS NOTE: All “client side” software natively bundled into most all distributions of LINUX NFSv3 NFSv3 NFSv3 NFSv4 NFSv4 NFSv4 pNFS pNFS pNFS Gluster Client Gluster Client Gluster Client NFS RDMA NFS RDMA NFS RDMA Spark Client Spark Client Spark Client Hadoop Client Hadoop Client Hadoop Client NFSv3 NFSv4 pNFS Gluster Client Hadoop Client Client Side Protocols: •  NFSv3 •  NFSv4 •  NFS RDMA •  pNFS •  Gluster Client •  Spark Client •  Hadoop Client 1. All clients can be the same protocol 2. Potential to intermix client protocols The most FLEXIBLE HyperParallel File System in the world!

Single Mount Point // PavilionPHFS FILE - Per Chassis: Read BW = 75+ GB/s Write BW = 50+ GB/s ( 67% of Read) Flexibility to Scale “in the chassis” + “cross chassis”  At linear performance 1 2 3 4 18.75+ READ 12.5+ WRITE 2 500TB 18.75+ READ 12.5+ WRITE 3 500TB 18.75+ READ 12.5+ WRITE 4 500TB 56.25+ GBps READ 37.5+ GBps WRITE 1.5PB 37.5+ GBps READ 25+ GBps WRITE 1PB 18.75+ GBps READ 12.5+ GBps WRITE 0.5PB 75+ GB/s READ 50+ GB/s WRITE 2PB 75+ Read 50+ Write 1 2PB 75+ GBps READ 50+ GBps WRITE 2PB 150+ GBps READ 100+ GBps WRITE 4PB 225+ GBps READ 150+ GBps WRITE 6PB 300+ GBps READ 200+ GBps WRITE 8PB 375+ GBps READ 250+ GBps WRITE 10PB 525+ GBps READ 350+ GBps WRITE 14PB 450+ GBps READ 300+ GBps WRITE 12PB 75+ Read 50+ Write 2 2PB 18.75+ READ 12.5+ WRITE 1 500TB 75+ Read 50+ Write 3 2PB 75+ Read 50+ Write 4 2PB 75+ Read 50+ Write 5 2PB 75+ Read 50+ Write 6 2PB 75+ Read 50+ Write 7 2PB 75+ Read 50+ Write 8 2PB 600+ GBps READ 400+ GBps WRITE 16PB Single Mount Point //PavilionPHFS UNLMITED scalability = Single Mount Point //PavilionPHFS Scale INSIDE the Chassis Scale OUTSIDE the Chassis (example 8 chassis – scale is unlimited) Best in Every Class Performance & Performance Density 600+ GB/s READ 400+ GB/s WRITE 16PB Usable 32 Rack Units OBJECT - Per Chassis: Read BW =50+ GB/s Write BW = 30+ GB/s ( 60% of Read) PLUS OBJECT Performance 400+ GB/s READ 240+ GB/s WRITE 16PB Usable 32 Rack Units

A Peek into the Software – Its power and simplicity

It’s Real (Product Demo) Sundar Kanthadi - Vice President of Engineering & Co-Founder Customer Engagement With The Product

Open Q&A Ask Away!

Unrivaled Choice and Control Amy Love - Chief Marketing Oﬃcer Positioning: Market Leadership Customer Choice & Control

Our Marketing Guiding Principles 36 WIN HEARTS, MINDS, & DOLLARS.   Enable sales. Drive revenue. THE BUYER’S JOURNEY Digital & analog THE BRAND IS A RELATIONSHIP The totality of who we are. Lead with strategy. " Tell magical stories. " Build immersive experiences. Analysts

Clear Conviction on Pavilion’s Value Proposition to Customers

Unrivaled Customer Choice File Object Block NFS v4 NFS RDMA NVMe-RoCE NVMe-TCP NVMe-RDMA NFS v3 iSCSI S3 pNFS Plug-ins NVMe-TCP for Windows

Unrivaled Customer Control NFS S3 S3 N F S NFS N F S

2021 Marketing Initiatives SIMPLIFY:  GTM & Enable Sales MARKETING INITIATIVE:  Visibility at Points of Self-Directed Discovery Thought leadership, brand placement, media and analyst engagement   Directed & Leveraged Visibility DRIVE:  Thought Leadership &   Accelerate our Brand GROW:  Demand & Revenue MARKETING INITIATIVE:  Contacts for Self-Directed Education and Comparison Nurtures, MDR, AE engagements Owned/controlled assets (website, social) EMPOWER:  Marketing Team Brand Drives Demand MARKETING INITIATIVE:  Continual honement of value-based messaging Speciﬁc Sales Plays  Competitive Insights     Content to drive inbound activity and prospect discussions MARKETING INITIATIVE:  Culture of test & trial, with continual exploration and adjustment Ongoing evaluation of spend eﬀectiveness and adjustments

2021 Marketing Actions & Investment Focus SIMPLIFY:  GTM & Enable Sales Key Programs Key Programs Key Programs DRIVE:  Thought Leadership &   Transform our Brand GROW:  Demand & Revenue Key Programs Content: Rationalize, focus, revitalize - deep vs. broad Playbooks: Clarity of action: aircover and ground assault to support channel and sales teams (MVP) Messaging & Positioning: SKO, Training POCs/Testing: Tell the stories of outcomes, virtual and On-prem, benchmarks Sales Enablement Tools: Sales plays: videos, customer stories, competitive insights, call decks, customer facing material, e-books, competitive, etc.; win/ loss reports MDR: Liaison: AEs, Mktg, Prospects EMPOWER:  Marketing Team and Broader GTM Team Website & Messaging: Superior customer outcomes/results Stories: Demonstration of outcomes Experiences/Events: Virtual - webinars, speaking engagements, and personalization Thought Leadership Campaigns: " 1H Universally Unmatched, Unrivaled Choice and Control Brand: Consistent visibility at destination sites, retargeting to maintain top of mind Advocacy: Customer stories/case studies References, Testimonials, Awards Influencers: Media, analysts, social Follow the Trail: Retargeting People: Team mission & vision, goals & metrics Website: Re-due (content and capabilities) Data & Infrastructure: Attribution capabilities Culture: Right Stuff, in it to win it, empowerment, action-orientation, accountability, best intentions, benefit of the doubt, fail fast Process: For scaling GTM/Launches: Programs and Kits to support effective execution and acceleration Sponsorships: Support AEs locally MarTech Stack: Digital engagement tools Intent & Assessment Capture Customers actively researching topics and comparing storage vendors Inbound Capture Social, influencer, brand Revenue Marketing: Acquire-Grow-Build strategic development of existing accounts Personas: Buying group who and why Account Based Marketing: TechTarget, ITCentral, Terminus, Alyce Channel/Partner Leverage: Education, joint outbound campaigns, press, & events, Nvidia, Carahsoft, Penguin Software Developer Leverage: Clarifi, OmniSci Select Investments:" PR firm/PR wire: Media relations" TechTarget: Retargeting" SEO/SEM: Retargeting" HPCwire: Industry publications" STAC/NAB/SC/FED: Events & Experiences" Evangelist: Thought leadership, visibility" " Select Investments:  MarTech Stack: Digital/Demand Gen Engine  TechTarget: Education Project Intent  ITCentralStation: Assessment Capture  ABM Programs: ABM" TechTarget: Display Adv., Content Syndication" Dir. Channel Mktg & Funds: Channel leverage" CAB/PAB: Relationship Expansion " " Select Investments:  Director Solutions Marketing: Content" MDR: AE Engagement" Customer Videos: Production content" Content Creator/Writer: Content" Analysts: Commissioned content, membership, on-the-record, virtual events" " Select Investments:  Website/Prgm Mgr: Overhaul" Design Firm or Person: Look & Feel" Programs: Resources for GTM and Launch campaigns" Sponsorships: AE leverage (local level)

Digital Presence

Thought Leadership, Media Engagements, Outlook and Predictions

Webinars Pavilion Webinars Presenter (P) Moderator (M) Additional Participant - Q&A February 4 Shatter IO Bottlenecks and Get More From Your Data Than You Ever Thought Possible Costa Hasapopoulos(P) Keith Parker (M) N/A February 25 Perhaps March How Sony is Changing Volumetric Capture & Accelerating Creative Outcomes Amy Love (M) Chris McBride (P) V.R. Satish, Pavilion Michael Pearce, Sony Billy Russell, Alliance IT March Accelerating GPU Based Applications with NVIDIA Validated Magnum IO GPUDirect Storage (With Carahsoft) Keith Ober (P) Keith Parker (M) Rob Davis, NVIDIA March Reduce cost and complexity while accelerating SQL Server applications Costa Hasapopoulos (P) Keith Parker (M) N/A Partner Webinars February 10 Accelerating GPU Based Applications with NVIDIA Validated Magnum IO GPUDirect Storage (With Carahsoft) Costa Hasapopoulos(P) Keith Parker (M) Rob Davis, NVIDIA March Scalable High Performance Storage for HPC (With Penguin) Costa Hasapopoulos (P) Amy Love (M) Kevin Tubbs, Penguin Computing

Universally Unmatched Solutions Costa Hasapopoulos - Chief Field Technology Oﬃcer & Worldwide VP Business Development Solutions Being Built With Partners (NVIDIA, OmniSci, Sony)

Fast NFS & Fast BLOCK – Many Uses Cases 46 Virtualized Infrastructures

New & Growing Partnerships Resell / Solution Partnerships Technology Partnerships

Michael Kagan – CTO and SC20 “High performance computing requires high performance IO” “The rule is simple, the higher processing power the computer element the more data it can process hence faster data delivery is required” Pavilion – “Performance improvement for file storage of almost 4x by using GPUDirect Storage” “Storage access is yet another bottleneck that needs to be resolved…”

Pavilion –  The Most Flexible High Performance Storage Platform File Object Block DENSITY Multi-Storage Type / Protocol Best in Every Class Performance NFS v3 NFS v4 NFS RDMA NVMe- RoCE NVMe-TCP NVMe-RDMA iSCSI S3 90 GB/s – Read 56 GB/s - Write Read 120 GB/s 20M IOPS @100 µsec Write 90 GB/s 5M IOPS @ 25 µsec Not just Performance Density But Flexible Performance Density All at market competitive pricing 4 Rack Units 2.2PB PB RAW 3200 Watts 20 Million IOPS 20 Controllers 40 100 Gb Ports 540TB per RU 5M IOPS per RU 800 Watts per RU 5 Controllers per RU 10 x 100Gb per RU File Object Block 80 GB/s – Read 35 GB/s - Write 191 GB/s @ 1.75ms READ 118 GB/s @ 5.6ms WRITE GDSIO-X0 DGX-A100 to 2 x Pavilion HFA’s 182 GB/s @ 3.87ms READ 149 GB/s @ 4.51ms WRITE GDSIO-X0 DGX-A100 to 2 x Pavilion HFA’s Note: All numbers were tested with ONE Pavilion chassis and 4 GPU’s and extrapolated to TWO Pavilion chassis (8RU) and 8 GPUS solution is expected to LINEARLY scale) “High performance computing requires high performance IO”

Pavilion’s NVIDIA Advantage….   Unlock the storage bottleneck and unleash the beast! Supports Block, File and/or Object File Object Block Supports GPUDirect Storage Block File OR Open Standards Based Protocols •  Ethernet and/or InfiniBand •  No Host Agents required •  NFS, NFS RDMA, NVMe-TCP, NVMe-RoCE, NVMe- RDMA, S3 Super High Performance •  Ultra Low Latency •  Impressive READ performance •  Equally impressive WRITE performance “Storage access is yet another bottleneck that needs to be resolved” Performance is “similar” to INTERNAL SSD’s… •  Enabled NVME-oF and RDMA technologies •  High speed ethernet and/or InfiniBand connections 4 Rack Units & up to 2PB Usable

Pavilion + DGX-A100 – A Powerful Combination – FILE or Block 51 70 GiB/s (75 GB/s) 178 GiB/s (191 GB/s) 110 GiB/s (118 GB/s) 69GiB/s (74 GB/s) 9.0ms 1.75ms 4.4ms 5.6ms 26% 16% 28% 11% BLOCK 69 GiB/s (74 GB/s) 170 GiB/s (182 GB/s) 139 GiB/s (149 GB/s) 70GiB/s (75 GB/s) 8.96ms 3.87ms 9.0ms 4.51ms 14% 4.4% 51% 4% Supporting NVIDIA GPUDirect Storage FILE or Block Dense Form Factor – 8 RU (4RU per Chassis) Ultra LOW Latency Impressive WRITE performance Notes: •  All IO tested using GDSIO and loads directly to the GPU (GPU Direct) or to the GPU’s via the CPU (NON GPU Direct) – this is NOT an FIO test •  % in YELLOW = CPU Utilization •  ms = Latency •  All numbers were tested with ONE Pavilion chassis and 4 GPU’s and extrapolated to TWO Pavilion chassis (8RU) and 8 GPUS solution is expected to LINEARLY scale) •  Two chassis were optimized for READ throughput to saturate the DGX-A100. We could add more chassis to saturate the WRITE throughput (not a focus on these tests) NO host agents required FILE 2.5x greater BW 2.5x lower Latency 1.8x lower CPU util 2.5x greater BW 2.3x lower Latency 11.5x lower CPU util GPU Direct NON GPU Direct GPU Direct NON GPU Direct

Spectrum Scale (GPFS) Solution – Government Customer World’s largest NVMe-oF Implementation Shared Volumes Very large scale-out file system Nearest competitor: •  40% our READ •  60% our WRITE •  ~150% our RU (per chassis) •  ~3 x 6 RU to our 1x4 RU “like for like” Spectrum Scale Performance: •  3 Pavilion “cluster” •  256GB/s READ •  220GB/s WRITE •  Single Pavilion •  85GB/s READ •  73GB/s WRITE Quote from partner: “Pavilion makes the impossible possible” Future… •  15 Pavilion 4RU boxes •  1000 to 2000 servers •  Fyi… $75k (4TB memory servers) •  EVERY server connecting NVMe-oF to EVERY pavilion box

Modern Private Cloud for Virtualizing Modern Apps  Requires consistent predictable performance high performance with low latency Huge consolidation play for VMware •  16 boxes to 1.5 Pavilion boxes •  NFS today moving to NVMe-oF •  NVMe-oF provides: •  3X the 4k Random Read •  4X the 4k Random Write •  2X Seq Read •  4.5X Seq Write •  22% Read Latency Reduction •  44% Write Latency Reduction Unique Benefits for Virtualized Infrastructures •  Performance of DAS benefits of SAN •  Scale Capacity independent of Processing power •  Multi-dimensional performance advantage = FLEXIBLITY •  High IOPS •  Low LATENCY •  High THROUGHPUT •  Virtual infrastructures require consistent predictable performance (i.e. original reason for DAS) VMware •  vCenter PLUG-IN •  NVMe-oF 7.0 certification Kubernetes •  CSI Driver Docker Container VMware VM NVMe & NVMe-oF NET NET More WORK being done on the SAME physical infrastructure!

Sony & Other Media and Entertainment Use Cases Media & Entertainment Windows NVMe-oF Initial Use Case Volumetric Capture •  Multi-Tiered Spectrum Scale Solution •  Many other uses cases under investigation Other Media and Entertainment Opportunities •  Post Production workflow •  Transcode / Encode solutions •  General TIERED storage solutions leveraging Spectrum Scale and others Windows NVMe-oF (low latency)

Big & Fast – SQL Server Acceleration Direct Attached LOW latency HIGH IOPs SQL Server Solution SQL Server 2 SQL 2 SQL 2 SQL Server 3 SQL 3 SQL 3 SQL Server 4 SQL 4 SQL 4 Driven by need to SCALE SQL Server – consistently and predictably •  Enabled by Windows NVMe-oF (low latency – high bandwidth) •  Legacy DAS (performance reasons) •  Move to SAN for flexibility and consistency •  Enabled by predictable consistent performance Financial Sector Firms •  Hedge Fund •  Service Provider Start small grow big… SQL Server 1 SQL 1 SQL 1

Looking into the near term future – NEW USE cases … High Performance Parallel File System High Performance Analytics Environments Medical Research

Looking into the near term future – NEW USE cases … High Performance Parallel File System High Performance Analytics Environments Multi-Sensor Data Analysis

Looking into the near term future – NEW USE cases … High Performance Parallel File System High Performance Analytics Environments Financial Analysis Multi-input, Multi-Format, Multi-User big data inputs and processing solutions driving the demand for a high performance scale out file system

Fast FILE, OBJECT, BLOCK – Many Uses Cases 59 Applications run faster, perform more consistently and scale larger To provide solutions for… And deliver cost effective outcomes to the business… High speed sensor ingest High frequency trading Machine Learning Artificial Intelligence Data Analytics Virtual Infrastructure Log analysis Product Development / Engineering File sharing Machine Learning GPU’s Enable Applications like…

Go To Market Success Chris McBride - Chief Revenue Oﬃcer Unmatched & Clear Separation in Analytics, VMware, M&E (LF/LB)

Our Market Focus Where it Matters Most o  AI/ML and Analytics o  Financial Services o  Cybersecurity o  Facial Recognition o  Signals Processing o  Extreme Science Discovery o  M&E Production/Postproduction o  Life Sciences o  High Performance Virtualized Infrastructure Key Differentiation o  Best Performance •  Writes •  Latency •  Reads o  Density – Performance/Density o  Flexibility o  Easy Scalability

Key Wins Common Themes o  The need for fast writes for data ingest coupled with high performance AI and Analytics •  Financial data feeds for Quant/Algorithmic Trading •  Images, Sensors, Signals Intelligence for Decision Support •  Virtual Production/Volumetric Capture o  Centralized high performance storage tier to support HPC workloads for AI and Analytics •  Academic Computing, Research, Life Sciences, Applied Physics •  Medical Billing Optimization o  High performance Analytics on Virtualized infrastructure •  Pavilion performance allowing customer to move Greenplum from bare metal to VMware

Key New Account Win VMware 7 over RoCE o  Committed to VMware o  Committed to moving to Converged Ethernet o  Looking for a storage platform to support high performance applications o  After testing: •  Pavilion enabled 50% increase in VM Density •  Resulted in 50% reduction in VMware and Oracle Licensee fees o  All expansion moving to Pavilion o  Will migrate 2PB VMware storage estate to Pavilion Source: RDMA Networking: Mellanox June 19, 2018

Case Studies and Use Cases Virtual Environments with SQL Server Facial Recognition Analytics Pavilion Resources Page

2021 Outlook o  Continue to expand interest and pipeline o  Strong bottom’s up visibility in support of 2021 objectives o  Most pipeline is tied to: o  AI/ML and Analytics Facial Recognition •  Cyber Security •  Signals Intelligence •  Life Sciences and Research •  Satellite/Geological Sensors •  Financial Services: Quant/Algorithmic Trading o  Media & Entertainment •  Volumetric Capture •  Virtual Production •  Editing and Voice Dubbing o  Exploring VMware 7 over RoCE interest •  Can deliver 50% reduction on apps licensed per socket

Open Q&A Ask Away!

Conclusion Gurpreet Singh - Chief Executive Oﬃcer On a mission!

Thank You! Send your contact information to [email protected] to be included on the bi-monthly analyst and inﬂuencer call and/or to schedule a follow on conversation.

IT Press Tour January 29, 2021

Pavilion Data - IT Press Tour Jan 2021

Pavilion Data - IT Press Tour Jan 2021

More Decks by The IT Press Tour

Other Decks in Technology

Featured

Transcript