9LivesData - IT Press Tour #65 Dec 2025

Transcript

1. Athens, Greece, December 9th, 2025
2. About the presenter Hi! I'm Cezary Dubnicki CEO @ 9LivesData from Poland 25+ years of experience in storage R&D Head of Storage Department NEC Labs Princeton, NJ PhD in Computer Science Google me and follow on LinkedIn I’d like to make you excited about 2
3. What is high9stor TCO-optimized Scale-out Enterprise Backup Storage 3
4. Key points • Market segment: Enterprise secondary storage (backup, archive) • User pain points addressed: Skyrocketing cost of backup infrastructure due to exponential data growth • What makes us different: A scale-out solution with TCO focus • Goals: 20% lower TCO today; 30% soon 4
5. 9livesdatamission and vision Mission Deliver efficient scale-out secondary storage built on commodity hardware to keep all enterprise backups secure and highly available while delivering low TCO of the entire system Vision Become a recognized player in delivering efficient solutions for the secondary storage market segment; Work towards standardization of interactions between backup and archiving applications and secondary storage 5
6. 9livesdata - companybehind the product oSpin-off from NECLA: NEC Labs America in Princeton NJ, USA o 9livesdata founder was head of NECLA Storage Department o Multiple 9livesdata team members consulted for NECLA o 9ld created to deliver NEC HYDRAstor oMain contract: NEC HYDRAstor o The first scale-out backup storage with global dedup o Significant scalability 165 2u nodes, 27 PB raw o For many years the fastest and the most scalable o Thousands of nodes deployed o Exabytes of data stored without any data loss with 16+ years experience in backup storage 6
7. NEC HYDRAstor work gave us enormous experience oDesigned the architecture and key algorithms o 24 US patents o 12 research papers in the best storage conferences/journals oImplemented 70% of the product code o Most complex components like distributed backend oHigh code quality o Bug density lower than major tech firms oProvided world-wide 3rd and 4th line of support o 9LivesData team delivers exceptional service and demonstrates a genuine commitment to customer satisfaction – a manager of a large bank in EMEA 7
8. Extreme real-life support story 1. Thursday: gas fire suppression system triggered in Data Center • Shock caused many HDD failures • Micro-vibrations led to severe data inconsistencies • 16-node cluster failed to start (75% HDDs with corrupted data) 2. Friday: client prepared for full system reinstall, total data loss 3. Weekend: data salvaged from damaged disks by 9livesdata team 4. Monday: production backups working again 5. Outcome: • Zero data loss for our system • Other systems suffered significant data loss 8
9. high9stor –a new 9livesdata product o9livesdata introducing a new product called high9stor o Scale-out secondary storage with global deduplication o Compatible with NEC HYDRAstor (our nodes can be used in NEC grids) o Please see www.high9stor.com for full feature set ohigh9stor as a software-defined storage o 1u SUPERMICRO nodes o High-density: 12x20 HDDs , 240 TB/u o Scalability: up to 180 nodes, 43 PB raw o Multiple server platforms planned 9
10. Customer challenges • How to deliver low TCO • How to meet backup window and restore targets in spite of exponential data growth • How to support backup applications • How to ensure business continuity • How to avoid capacity / performance bottlenecks • How to ensure high data availability • How to support non-stop operation 10
11. How to deliver low TCO ✓Reduced footprint due to high-density ✓ 240 TB/u with cheap SATA drives ✓Very efficient storage usage ✓ Global dedup and compression ✓ High storage utilization supported ✓ Dynamic sharing of capacity ✓ Tiering to cloud ✓Long-lived hardware nodes ✓ Older nodes used with newer gen nodes ✓Significant savings on management ✓ One pool easy to manage ✓Sourcing from multiple hardware vendors 11 $ $ $
12. How to meet backup window and restore targets ✓Scale-out architecture ✓ Performance grows with each node added ✓Dynamic division of performance among all clients ✓All disks in the system used for a restore ✓Extreme backup performance ✓ Up to 72 TB/h/node for writing with dedup on a client ✓Multiple agents allowing dedup on the client ✓ File-system compatible general protocol ✓ Specialized agents like OST plugin 12 DATA DELUGE (DATA WAVE) 11001011011010111111110011 1101110010111101111100101 10011001110110110110010 011100001111011111001 10011010111111111110 10001001110111110 100101101011111
13. How to support existing backup applications ✓One scale-out system accepting data from multitude of backup apps ✓ Global dedup for all data stored ✓Support for standard interfaces ✓ CIFS, NFS, S3 ✓ Dedup on client using file system interface ✓Tight integration with selected applications ✓ NetBackup OST ✓ VEEAM Fast Clone, Object Lock ✓ COMMVAULT Object Lock ✓ NAKIVO optimized synthetic ✓ And others ... 13
14. Ransomware protection / disaster recovery ✓WORM integrated with backup applications ✓ NetBackup OST integration ✓ S3/Object lock integration – VEEAM, COMMVAULT and others ✓ File-sys based for COMMVAULT, NAKIVO ✓Multisite, dedup-optimized replication ✓ Integrated with backup apps ✓ Sends only unique data ✓ In-flight encryption ✓Instant recovery support 14 Protection period configurable at file level WORM: Write Once Read Many 5 years protection 10 years protection
15. How to avoid capacity and performance bottlenecks ✓One scale-out capacity and performance pool ✓ Dynamically shared across all clients ✓ Adding a node increases both capacity and performance ✓In-line global dedup ✓ Dedup & compression save a lot of capacity ✓ Dedup makes backups several times faster ✓Very fast background tasks ✓ Multiple controllers (nodes) working together ✓ Only few hours to reclaim space from deleted data ✓ For some competitors it may take weeks ✓Consistent operation with very high capacity utilization 15 B C A C B A C B F D E D E D F F Duplicate elimination 1 A B C D E F Compression A B C D E F In-line block-based dedup
16. High-performance background tasks 16 high9stor: scale-out architecture Each node performs tasks on it's own capacity, so total capacity growth doesn't extend duration Many workers Competition with scale-up architecture A big controller does all tasks, So the duration of tasks grows with capacity growth Single worker Customer case: Reduced the time to reclaim storage capacity 67 times - from 2 weeks to 5 hours!
17. How to ensure data availability 17 high9stor High data availability and resiliency due to erasure codes Up to 6 node/disk failures tolerated per 12 disks/nodes Data chunk 1 4 7 2 5 8 3 6 9 10 9 original subchunks 3 encoded subchunks 1 7 11 9 2 12 10 6 4 5 8 3 nodes 11 12 Naive approach: replication 4- way replication needed to ensure 3 disk failure resiliency Data chunk Copy 1 Copy 2 Copy 3 Percent of storage used for resiliency 25% 75%
18. Non-stop operation 18 Need for on-line system growth due to increased volumes of protected data Need for on-line hardware refresh Data lives 7-10 years vs. Hardware 3-5 years Data automatically rebalanced after node addition/deletion v3 v2 v1 v3 v3 v2 v1 v1 v1 Nodes added/replaced without interrupting system operation Up to 3 generations of nodes can work in one system Avoiding fork-lift upgrades
19. Scale-out Architecture NSPOF grid with capacity/performance scaling high9stor selected features & differentiators 19 Data Durability Erasure Coding - up to 6 nodes/disks failures resiliency in groups of 12 Data Reduction Inline variable block global deduplication and compression Compliance Dynamic data shredding and In-flight & at-rest encryption Ransomware Protection WORM (write-once-read-many) integrated with backup applications Disaster Recovery WAN-Optimized remote replication Autonomous Operation Self-healing, on-line node addition and deletion Simplified Maintenance No forklift upgrade, up to 3 generations in one system Backup App Integration Optimized synthetics and copy (e.g. OST),source-side dedup, WORM Cloud Tiering Tiering of older backups to public clouds (only unique deduped data) Rich Interfaces Standard NFS/CIFS/S3, Backup-app specific OST
20. Comparison with competition 20 Competitor Software Defined or Appliance Architect ure Max Raw Capacity PB HDD used TB Storage Density TB/u Dedup 9ld high9stor Software Defined Scale-out 43 (180 1u nodes) 20 240 inline Dell DataDomain Appliance Scale-up 1.5 (7 shelves) 8 77 inline ExaGrid Appliance Scale-out 13.8 (32 2u nodes) 18 216 landing zone (no dedup) + offline HPE StoreOnce Both Scale- up/out 1.7 8 77 inline Cohesity Both Scale-out Petabytes, 2u nodes 24 144 inline Rubrik Both Scale-out Petabytes 2u nodes 16 91 inline
21. Competitive differentiation ✓TCO reduction focus ✓High density compared to HDD-based competitors ✓ 240 TB/u ✓ Feature requested by our clients ✓Global in-line dedup in a scale-out system ✓ Efficient storage usage ✓ Very high ingress speed ✓Very fast background tasks due to distributed multi-controller algorithms ✓Non disruptive generation upgrade ✓No need for a Landing Zone 21
22. high9stor as a Software Defined Storage (SDS) ✓9livesdata delivers high9stor software only ✓Compatible hardware defined by hardware compatibility list ✓ Hardware purchased separately by a system integrator ✓ Version, vendor, bill-of-materials ✓ All version nodes compatible ✓SDS shortens the development cycle ✓Software reuse in case of hardware changes ✓Possibility to follow market trends (HDD, SSD, CPU..) ✓Having multiple hardware suppliers reduces Capex 22
23. Pricing model ✓Recurring software license based on raw TB/month ✓ Effective capacity after erasure code overhead depends on user selection ✓ License price proportional to capacity x time ✓ Easy to understand for users ✓5 years or 3 years paid up-front ✓ One-year extensions afterwards ✓About 20% cheaper than competitive solution 23 Price ~ Time x Raw Capacity Time Raw Capacity
24. Partners ✓Data Sciences Corporation ✓Biggest NEC HYDRAstor distributor Africa, Middle East ✓User needs help define the product ✓Large NEC HYDRAstor installations ✓Natural market for node replacement with high9stor nodes ✓Technology partners ✓SUPERMICRO, Backup SW vendors 24 Data Sciences Corporation
25. Case study –A big EMEA financial institution ✓4 grids in 2 data centers - hundreds of nodes in three generations ✓NetBackup/OST with replication between data centers ✓320 PB effective capacity 25 v3 v3 v2 v2 v1 v1 high9 stor high9 stor 1. Now: gradually replace older nodes with high9stor nodes 2. Early 2026: add new high9stor grids HYDRAstor Action plan
26. high9stor already stable Over the past several months, Data Sciences has evaluated 9LivesData’s high9stor scale-out backup storage platform. They demonstrated the same commitment to product quality that we observed over the last decade of HYDRAstor support. As a result, Data Sciences Corporation is confident in the competency, professionalism, and capability of 9LivesData. We look forward to continuing to expand our deployment footprint with them in the region. -- Gerard Almon, CEO of Data Sciences Corporation 26
27. Which geographic markets ➢Focus on EMEA and Central Asia excluding countries under sanctions 27
28. Target market segments ➢Enterprises ➢Financial institutions ➢Energy and Utilities ➢Telecoms ➢Media and Entertainment ➢Health care, hospitals 28 ➢Government ➢Made-in-EU advantage ➢Social security ➢National health ➢Finance and taxation ➢Motor vehicles DBs
29. How to sell ➢Selling via integrators and sales agents ➢Sales agents close deals only ➢9livesdata acts as a system integrator ➢Use outsourcing as much as possible … ➢Break-fix ➢First lines of support ➢But, one point of contact for a customer Service Integrators Sales agents Close deals Integrator Sales agent Break-fix Integrator Outsourced Installation Integrator 9livesdata Basic support Integrator Outsourced Advanced support 9livesdata 9livesdata 29
30. Roadmap ➢Modify existing systems to save up to 50% of power consumption ➢Deliver S3 dedup on client ➢Release archive version ➢Certify with additional applications (Backup and Archive SW) ➢Support for multi-actuator drives once larger capacities available 30
31. Future directions ➢Deliver next gen high9stor utilizing QLC SSDs ➢ Much higher performance and very high density ➢ TCO comparable to the current high9stor despite using SSDs ➢ More environment friendly ➢Work with backup app vendors to standardize interaction between backup app and storage, for example ➢ Metadata-only copy and full backup synthesis ➢ Fully integrated WORM, replication 31
32. Thank you for listening 32 Thank you, Cezary Dubnicki CEO, 9LivesData dubnicki@9livesdata.com https://www.linkedin.com/in/cezary-dubnicki/ Visit our website high9stor.com