StorageReview.com

By StorageReview.comTechnology

StorageReview.com episodes

September 29Podcast #153: ServerPartDeals CEO Max Manekia on Refurbished Drives as Supply Chain Strategy
Brian recently sat down with Max Manekia, CEO of ServerPartDeals, at the company’s facility in Sanford, Florida, on the same trip that produced our ServerPartDeals facility tour. We’ve been a ServerPartDeals customer for years. Brian put them on our drive leaderboard not long after they opened their doors in 2010, and our 2025 look at recertified hard drives was built on their inventory. This conversation covers how a home lab supplier ended up in enterprise procurement plans, and what it takes to test, grade, and ship as many as 2,200 drives a day.
The ServerPartDeals StorNado test wall, which runs 24 hours a day
Max Manekia is the CEO of ServerPartDeals, the Sanford, Florida reseller of recertified enterprise hard drives, SSDs, servers, GPUs, and memory. The company has served home lab builders since 2010 and now supplies enterprise buyers by the pallet, with more than 200 unique drive SKUs and tens of thousands of units in stock at any given time.
The compute shortage has become a storage shortage, and this podcast is a look at what the refurbished supply chain looks like when it’s run at scale: sourcing, 100% sector scans, grading, pricing, advance RMA, custom foam, and procurement as a service for buyers who need drives every month. It runs about 38 minutes, and Max is direct about what works and what still might need another expansion.
If you don’t have time to watch this end-to-end, we have broken it into segments so you can choose what’s most important to you.
[00:00] From Home Lab Staple to Supply Chain Player
The opening segment frames why refurbished enterprise drives have moved from a home lab budget play into mainstream procurement, and why ServerPartDeals’ early focus on reliability set it up for that shift.
The compute shortage has triggered a downstream storage shortage, and refurbished components are now part of mainstream procurement strategies.
ServerPartDeals has served the home lab market for more than a decade and is now selling to a more enterprise-oriented customer base.
An obsessive focus on product reliability from day one turned out to be the right foundation for where the market went.
A 42% repeat-customer rate reflects consistent supply, competitive pricing, and product quality customers can trust.
Customers are buying new drives from ServerPartDeals and sending their old drives back for recertification, which creates a circular supply loop and turns customers into suppliers.
[02:47] Receiving, Sourcing, and Saying No to Bad Inventory
This section walks through what happens before and after a case of drives lands on the receiving dock, and how the buying team keeps bad inventory from entering the pipeline.
Every inbound shipment is vetted before it arrives, with buyers reviewing provenance, firmware, date stamps, and supplier history up front.
Physical inspection at receiving confirms part numbers and cosmetic condition and flags anything that would disqualify a drive from sale.
ServerPartDeals sources from hundreds of relationships built over years, which keeps inventory flowing without the volatility of spot buys.
They were among the first online resellers to post real product photos in place of stock renders, an early standard for transparency.
Testing queues are actively managed, with priority given to drives already sold to customers on the waiting list.
[10:26] The Storinator Test Wall
The testing floor is the heart of the operation. This segment covers how long a full scan takes, what throughput looks like today, and the power problem that came with growth.
High-capacity drives get 24- to 36-hour full-sector scans, and current throughput runs 1,800 to 2,200 units per day.
A 20TB drive alone needs 30-plus hours on the test rigs, which run around the clock to keep pace with demand.
A year ago the goal was 1,000 drives a day. The team hit 2,000, and the next constraint may be the building itself.
Power has been the hard limit. The facility used its existing three-phase, 600-amp service and drew additional capacity from a neighboring tenant.
100% sector scans are the single biggest reducer of RMAs, which is why every drive gets one before it is listed.
[14:20] Grading, Provenance, and Fraud
There is no industry-standard grading system for used drives, so ServerPartDeals defines its own. This section explains the grades and how the company guards against counterfeit and misrepresented products.
Grade A means zero defects. Grade B and C products are sold wholesale and never to end customers on the main storefront.
Drives with bad sectors still have value to the right buyer, but they carry no warranty and are not listed on the site.
Fraudulent product is a known risk in the secondary market. New suppliers have to send samples first, and every drive is checked between purchase and live scan.
[20:00] From Scan-In to Live Listing, and How Pricing Works
Once a drive passes, it goes live across every channel automatically. This segment covers the listing pipeline and the pricing model behind it.
Graded drives are scanned into live inventory and listed on ServerPartDeals.com, eBay, Amazon, and Newegg with pricing, specs, photos, and warranty details.
Pricing is managed as a weighted average against the broader market, factoring in warranty length, availability, and competing options. The goal is not the lowest price on the internet.
The catalog runs to more than 200 unique SKUs with tens of thousands of drives in stock, and a typical order runs $1,000 or more.
[22:34] Warranty and the RMA Cage
Warranty claims are where the front-end testing pays off. This section explains the advance replacement process and why the RMA area is the smallest room in the building.
Warranty claims are usually handled with advance replacements. A screenshot of the failure is enough to get a replacement shipped before the original comes back.
The RMA area occupies roughly 120 square feet, a footprint that reflects how much the sector scans catch before drives reach customers.
[27:15] Shipping and the Foam
Packing is engineered, not improvised. This segment covers the custom foam inserts and the multi-drive trays that replaced stacks of single boxes.
Custom foam inserts engineered to a 200G drop rating protect individual drives in transit.
Multi-drive foam trays were developed separately to handle orders of three to ten units without stacking individual boxes.
[32:30] Two-Day Air, Procurement as a Service, and the Full Stack
The closing segment covers how orders move out the door, the account management model for recurring buyers, and how far beyond hard drives the catalog now reaches.
Orders ship primarily via two-day air. The stated cutoff is 3:30 p.m., but the team keeps processing through 6 p.m., and many orders arrive next day.
Procurement as a service pairs recurring buyers with dedicated account managers who plan monthly orders and lock in consistent supply.
ServerPartDeals covers the full stack beyond hard drives, including servers, GPUs, RAM, and NVMe SSDs, and its sourcing relationships can sometimes beat standard lead times.
A new “Can’t Find It” button on the site lets customers request unlisted inventory, and a team member follows up within 24 hours to source or quote the item.
To keep up with ServerPartDeals and the refurbished drive market, follow Max Manekia on LinkedIn.
This podcast is sponsored by ServerPartDeals. All views and opinions expressed in this podcast are based on our unbiased view of the product(s) and services under consideration.
The post Podcast #153: ServerPartDeals CEO Max Manekia on Refurbished Drives as Supply Chain Strategy appeared first on StorageReview.com.
38 min
August 27Podcast #152: Data Center Thermal Dynamics and Cooling Architectures With Tim Shedd
Brian recently caught up with Tim Shedd, an expert in thermal management and a pioneer in liquid cooling technologies. The podcast covers some of the historical aspects of data center liquid cooling, from Tim’s time working with Cray and spraying liquid across tubes to his more recent engagement with Dell’s PowerEdge XE9680.

Dell PowerEdge XE9680 in the StorageReview lab
Tim Shedd is an industry expert and mechanical engineer specializing in thermal management, liquid cooling, and data center infrastructure. With a career spanning academic research as a professor at the University of Wisconsin-Madison, foundational engineering roles at Motivair, and senior technical leadership within Dell Technologies’ CTO organization, he has been at the forefront of high-density cooling architecture. His work focuses on bridging advanced thermal physics and hyperscale manufacturability, establishing standards for cooling distribution units, and optimizing data center power efficiency across air- and liquid-cooled hardware topologies.
This podcast delivers great insight into the evolution of thermals from mainframes, HPC platforms, and current server technology. In under an hour, you will come away with a better understanding of how this affects many aspects of keeping our technologies running efficiently.
If you don’t have time to watch this end-to-end, we have broken it into five-minute segments so you can choose what’s most important to you.
[00:00] The Evolution of Liquid Cooling from HPC to Enterprise AI
The initial segment outlines the historical progression of liquid cooling from specialized supercomputing installations to high-density commercial enterprise applications, highlighting early thermal design choices.
Early compute environments relied on evaporative cooling and basic forced-air convection before reaching the limits imposed by high-density HPC and accelerated compute workloads.
Initial academic experimentation with Cray architectures involved direct liquid spray over tubes, proving technically viable but commercially cost-prohibitive.
The post-2015 inflection in rack power density pushed OEMs to engineer standardized enterprise liquid deployments, culminating in high-density rack production models by 2019.
OEM supply chains pivoted from producing a few thousand liquid-cooled nodes annually to scaling delivery into thousands of production systems weekly.
Workhorse enterprise platforms like the PowerEdge XE9680, the fastest-growing server in Dell history, balanced early AI thermal loads using dense air heatsinks paired with rear-door heat exchangers before full direct-to-chip adoption became mandatory.
[05:00] Testing Thermal Breakpoints and Fluid Physics at Scale
This section explores the laboratory evaluation of competing thermal technologies, detailing why single-phase direct-to-chip loops emerged as the dominant architecture over immersion systems.
CoolIT direct-to-chip cold plates inside a Dell PowerEdge R760
OEM validation labs conducted extensive testing across cold-plate loops, negative-pressure networks, immersion configurations, and two-phase systems to establish deployment baselines.
Silicon thermal loads historically increased incrementally with each generation, allowing standard air-cooling and basic liquid-cooling approaches to suffice until TDPs exceeded the 500W-1000W range.
Immersion cooling offers operational advantages for broadly distributed thermal profiles, such as crypto farms, but struggles with the localized, highly concentrated heat fluxes found on modern accelerators.
Water blended with 25% propylene glycol (PG25), delivered through microchannel cold plates, became the primary standard due to its predictable scaling, supply availability, and established thermal transfer properties.
Open ecosystem standards remain critical for sourcing cold plates, manifolds, and quick disconnects across multiple vendors without performance regressions.
[10:00] Standardization, Multi-Vendor Interoperability, and Mechanical Reliability
The discussion transitions to the operational realities of deploying liquid hardware, focusing on multi-vendor component integration, liability demarcation, and material integrity.
Mixing cooling components from different vendors introduces significant mechanical compatibility hurdles and strict warranty limitations in the event of a fluid leak.
Legal liability and service-level agreements remain primary industry obstacles preventing mixed-vendor plumbing architectures across production server floors.
Legacy rope-style leak detection systems are giving way to integrated sensor topologies with automated pump trips and valve cutoffs.
Advancements in peroxide-cured EPDM hose manufacturing have significantly reduced component-level failure rates in direct-to-chip plumbing.
System unreliability in the field predominantly stems from manufacturing debris and inadequate initial line flushing rather than raw component fatigue.
[15:00] Debris Management, Heat Flux Limits, and the Physics of Negative Pressure
This timeframe examines precision cold-plate fluid dynamics, the catastrophic effects of microscopic particulate contamination, and the atmospheric physics that limit negative-pressure cooling loops.
Chilldyne negative-pressure CDU
Particulates of 100 to 200 microns, even a single wire-brush fiber caught in a quick disconnect, can obstruct microchannel fins and take down a rack through thermal throttling.
Advanced computational fluid dynamics (CFD) modeling enables modern cold plates to dissipate 1500W loads while maintaining a narrow 30°C delta T penalty.
Negative-pressure cooling systems are inherently constrained by atmospheric pressure, leaving only 7 to 9 PSI of usable differential pressure before the fluid reaches its boiling point at room temperature.
Limited pressure budgets make negative-pressure loops difficult to route through long facility manifolds, CDUs, and high-resistance microchannel cold plates.
Dedicated water chemistry analysis and ongoing chemical monitoring are essential to prevent corrosion, biological growth, and material degradation.
[20:00] Environmental Variables, Fluid Chemistry, and Advanced Leak Detection
Here, the conversation addresses regional climate impacts on data center operations, condensation risks, and the technological evolution of rapid-response leak-detection sensors.
Regional humidity, altitude, and ambient conditions dictate secondary-loop water temperatures and facility cooling efficiency.
Sub-ambient loop temperatures in high-humidity climates risk condensation formation on server chassis components and require dedicated dew-point controls.
Traditional leak detection ropes suffer from supply chain bottlenecks and high trigger thresholds, requiring a significant volume of fluid before triggering an alert.
Next-generation leak detection leverages flexible polymer-printed sensor traces, optical dye detection, and vapor sniffers to identify micro-leaks before catastrophic failure occurs.
Intelligent rack controllers actively modulate CDU loop pressure upon sensing early fluid loss to minimize leak volume while preserving uptime.
[25:00] Evaluating Two-Phase Direct-to-Chip Potential and Refrigerant Chemistry
This segment explores the thermodynamic benefits of latent heat vaporization in two-phase cooling, alongside the supply-chain and chemical considerations surrounding low-GWP refrigerants.
Two-phase direct-to-chip cooling eliminates the sensible heating penalty of single-phase loops, typically at least 5°C, maintaining a constant boiling temperature whether a chip dissipates 10W or 1,000W.
Two-phase cooling minimizes water-related corrosion and the risk of electrical shorting in the rack, presenting a compelling deployment model for enterprise data halls.
Scaling two-phase loops requires solving multi-vendor interoperability for phase-change manifolds, CDUs, and complex vapor condensers.
Emerging low-GWP, non-flammable dielectric fluids (such as R-515B and R-1233zd) are overcoming regulatory constraints associated with legacy PFAS compounds.
Modern data center infrastructure providers are acquiring specialist cooling vendors to assemble unified, turnkey liquid thermal portfolios.
[31:33] CDU Architectures, Dynamic Flow Control, and Blast Radius Mitigation
Focusing on the core pumping infrastructure, this section analyzes standardized testing methods, transient thermal response, and the engineering trade-offs between centralized and distributed CDUs.
Dell PowerCool CDU
ASHRAE Standard 127 established standardized testing methods to provide accurate, apples-to-apples comparisons of CDU thermal efficiency and pumping performance.
Minor variations in coolant flow translate immediately into temperature fluctuations at the cold plate; a 10 percent flow change can trigger throttling events, demanding ultra-responsive control algorithms.
Centralized multi-megawatt CDUs involve massive fluid volumes and large operational blast radii, such that a single contamination or leak event can take down an entire facility hall.
Additive chemical incompatibilities make mixing proprietary coolants from different vendors hazardous, potentially causing precipitation and clogging.
Distributed row-level and in-rack CDUs isolate mechanical failure domains, reduce fluid transit piping, and accelerate installation timelines.
[37:00] In-Rack CDUs versus Centralized Loops and the Viability of Air-Cooled Inference
This part evaluates high-performance in-rack CDU performance profiles and outlines why air cooling remains dominant across distributed enterprise inference clusters.
Purpose-built in-rack CDUs can support 220kW thermal loads with a 4°C approach temperature, turning 41°C facility water into 45°C coolant for Vera Rubin racks at 1.5 liters per kilowatt.
Dual-pump redundancy inside in-rack CDUs maintains continuous cooling operation even during single-pump maintenance events.
Air cooling accounts for roughly 90 percent of all global server shipments and supports a substantial portion of enterprise inference deployments.
Edge and enterprise inference hardware must be deployed directly within legacy, 10- to 15-year-old, air-cooled data halls where data is generated.
Retrofitting older facilities with rear-door heat exchangers and evaporative cooling towers shifts parasitic cooling power into usable compute kilowatts for local AI inference boxes.
[45:20] Chassis Internal Aerodynamics and Fan Efficiency Frontiers
The final segment breaks down the mechanical trade-offs of closed-loop all-in-one coolers and examines how chassis aerodynamic optimization drives air-cooling capability to its physical limit.
Server-level internal all-in-one liquid loops introduce mechanical failure points and chassis packaging complexity while offering only marginal gains over optimized vapor-chamber heatsinks.
Sidecar liquid-to-air heat exchangers provide a cleaner, more serviceable path for localized liquid-cooled racks in facilities lacking facility water.
Enterprise server fan efficiencies have increased from 15 to 20 percent a decade ago to modern electrical-to-mechanical conversion rates of 50 to 60 percent.
Optimizing chassis aerodynamics around a 100 CFM per kilowatt design point, with every bit of intake air warming a uniform 18°C, maximizes thermal dissipation per kilowatt.
Air has favorable thermal diffusivity properties that, when paired with precision ducting and advanced heat sinks, provide a solid foundation for next-generation enterprise hardware.
To keep up with everything around liquid cooling and efficient data center design, follow Tim Shedd on LinkedIn.
The post Podcast #152: Data Center Thermal Dynamics and Cooling Architectures With Tim Shedd appeared first on StorageReview.com.
0 min
July 13Podcast #150: HPE Alletra Storage MP B10000: Meeting Cyber Attacks Head-On
Brian connected with HPE’s Dan Gardner to follow up on HPE Discover and go into more detail on the HPE Alletra Storage MP B10000.

Dan is the Senior Technical Marketing Engineer at Hewlett Packard Enterprise and considers himself a security nerd. Dan joined HP in 2017 in hardware technical support, then moved into a technical support role in network and security. He moved to HPE in his current capacity in 2023. Dan is based in Birmingham, England, where, as he puts it, he would probably rather be up a mountain. He is also an ultra-runner.
Dan was instrumental in supporting our recent deep-dive coverage of the Alletra Storage MP B10000, making the conversation relaxed and informative. Following a brief overview of HPE Discover, Brian and Dan delve into the B10000’s internal workings. There is an obvious focus on cybersecurity, given the role the Alletra Storage MP B10000 and X10000 play in protecting enterprise data.
This is a great follow-up to our B10000 review published earlier this month. If you have 40 minutes, give this a watch or listen. If you are strapped for time, we have broken the podcast into five-minute sections below, so you can hop around to the topics that interest you.
There are several links after this transcript that provide more detail on the HPE Alletra Storage MP B10000, covering cyber resilience, ransomware, administrative tools, and more.
0:00 to 5:00 | HPE’s Unified Storage Vision and the Shift to Cyber Resilience
A look back at HPE Discover and why the B10000 and X10000 platforms were among the event’s biggest infrastructure announcements.
How HPE consolidated multiple storage platforms into a common hardware architecture that supports both structured and unstructured workloads.
The differences between the B10000 for block and file storage and the X10000 for object and file storage, while sharing the same scale-out foundation.
Why HPE believes cyber resilience has overtaken raw performance as the top priority for enterprise storage customers.
The discussion shifts from preventing ransomware attacks to recovering quickly with trusted restore points.
New legal and regulatory requirements are making cyber resilience a board-level business issue, not just an IT concern.
5:00 to 10:00 | Detecting Threats Earlier and Building Smarter Storage Security
Why enterprise databases and block storage remain some of the most valuable ransomware targets.
How HPE is working to bridge the gap between storage administrators and security operations teams.
The three pillars of HPE’s cyber resilience strategy: protect infrastructure, protect data, and detect threats in real time.
Moving beyond traditional ransomware indicators like exploding capacity usage to behavioral analytics.
Detecting suspicious activity, such as failed logins, privilege escalations, and snapshot deletion attempts, before encryption spreads.
An introduction to HPE’s ransomware detection capabilities built directly into the B10000 platform.
10:00 to 15:00 | Inside HPE’s Cybersecurity Lab and Continuous Ransomware Testing
An inside look at HPE’s Cybersecurity Center of Excellence and its isolated ransomware testing environment.
How engineers continually test against roughly 200 ransomware variants to improve detection algorithms.
Why ransomware detection is a constantly evolving process rather than a one-time feature.
What’s new in the latest B10000 software release, including improvements for file workloads and integrated data protection.
The role of HPE Morpheus and VM Essentials in helping organizations modernize virtualization alongside storage security.
Why HPE views cyber resilience as a full-stack challenge spanning compute, networking, virtualization, storage, and backup.
15:00 to 20:00 | Faster Recovery Through Automation and Security Hardening
How the B10000 automatically preserves snapshots when ransomware activity is detected.
New automation extends immutable snapshot retention to increase the chances of finding a clean recovery point.
Automated security hardening simplifies deployment of consistent security configurations across multiple arrays.
Built-in auditing continuously checks for configuration drift and alerts security teams when systems deviate from policy.
Why centralized audit logging is becoming essential for meeting modern compliance requirements.
An overview of emerging regulations, such as DORA and NIS2, and their impact on storage infrastructure.
20:00 to 25:00 | Practical Cybersecurity Advice for Organizations of Every Size
Why cyber resilience requirements are expanding well beyond finance and healthcare.
The growing influence of cyber insurance on backup, recovery testing, and operational readiness.
Practical recommendations for smaller IT teams looking to strengthen security without enterprise-sized budgets.
Why the NIST Cybersecurity Framework 2.0 provides a useful roadmap for organizations just getting started.
The importance of establishing a common cybersecurity language across executives, IT, security, and compliance teams.
Centralizing logs from every device becomes the foundation for meaningful threat detection.
25:00 to 30:00 | Building Visibility Across Hybrid Infrastructure
Open source tools such as Graylog provide accessible entry points for centralized security monitoring.
Why hybrid cloud environments require the same security discipline as traditional data centers.
Avoiding the assumption that cloud providers handle every aspect of security.
The importance of maintaining complete visibility across on-premises and cloud environments for governance and audits.
Security remains focused on protecting the entire attack surface, regardless of where workloads reside.
Discussion returns to how ongoing B10000 platform enhancements continue to improve scalability and resiliency.
30:00 to 35:00 | Security Fundamentals That Scale With the Business
Why basic security practices, such as monitoring privileged access, remain some of the most effective defenses.
The realities that smaller organizations face as infrastructure grows organically over time.
How B10000 delivers enterprise cyber resilience features even in entry-level deployments.
Organizations can expand protection over time with replication, backup appliances, immutable storage, and cyber vaults.
The concept of a “vault in a box” combines encryption, immutability, and integrated security in a single platform.
Good security hygiene today creates the foundation for future AI initiatives and regulatory compliance.
35:00 to 40:30 | Disaster Recovery, Performance, and Final Takeaways
Revisiting the 3-2-1-1 backup strategy and how B10000 supports modern disaster recovery architectures.
Using Metro Cluster, Active Persistence, remote replication, and StoreOnce to build resilient multi-site environments.
Why restore performance has become just as important as backup performance.
High-speed recovery places new demands on server, storage, and networking infrastructure.
Additional HPE resources, white papers, architecture guides, and demonstration videos are available for deeper learning.
Closing thoughts on why organizations of any size can benefit from adopting a more comprehensive cyber resilience strategy.
Additional Material
HPE Alletra Storage MP B10000 – Cyber resilience with data-adaptive ransomware detection
HPE Alletra Storage MP B10000 ransomware protection framework
HPE Alletra Storage MP B10000 SIEM integration
HPE Alletra Storage MP B10000 – Administrative hardening guide
HPE Alletra Storage MP B10000 security guide
Videos
HPE Alletra Storage MP B10000 ransomware detection with CrowdStrike
HPE Alletra Storage MP B10000 and Elastic
The post Podcast #150: HPE Alletra Storage MP B10000: Meeting Cyber Attacks Head-On appeared first on StorageReview.com.
41 min
May 29Podcast #149: Dell PowerStore Elite: Hands-on Deep Dive From the Hopkinton Lab
The Storage Review team visited the Dell campus in Hopkinton to go hands-on with the latest advancements in enterprise storage. This podcast takes place in one of our favorite places, the Hopkinton Dell Lab. Brian connects with Scott Delandy, an Engineering Technologist at Dell, for this interaction to go in-depth on the PowerStore Elite. This discussion covers the transition from PowerStore Gen 2 to the new PowerStore Elite, exploring the hardware architectural shifts and the software-driven intelligence defined by PowerStore OS 5.0.
While in the Hopkinton lab, we filmed a deep dive with Dell’s Technical Marketing Engineer, Stephen Granger, where Kevin and Steve tear down the 9500 controller and compare it to the 1500 controller. It gets technical with details around PCIe lanes, drive connectivity, battery backup architecture, serviceability, and new IO module form factor. It’s worth a look as a companion to the podcast video.
This short podcast and the Hardware Deep Dive videos literally tear down the PowerStore Gen Elite to give you an in-depth view of the internals. Brian and the team cover the upgrades from Gen 2 to Gen 3, providing the information you need to understand how the PowerStore Elite fits your environment. If you don’t have 30 minutes to spend, we’ve broken the segments down into five-minute increments so you can hop around to the sections that apply to your environment and interests.
[00:00-05:00] Introduction
The conversation opens with a transition from lab testing to the platform’s official rebranding and the key components of the new release.
Official rebranding of PowerStore Gen 3 to PowerStore Elite.
Integration of PowerStore OS 5.0 with the new Gen 3 hardware platform.
Backward compatibility of OS 5.0 features for existing PowerStore installations.
Hardware architecture leap featuring PCIe Gen 5 and DDR5 memory.
Shift to E3.S and E3.L flash media modules for high-density storage.
Standardization on OCP IO modules to streamline the component ecosystem.
[05:00-10:00] Hardware Standardization and Strategy
The dialogue focuses on the strategic decision to leverage common components across the Dell ISG portfolio and the benefits for supply chain resiliency.
Supply chain advantages gained by using standardized PowerEdge server components.
Reduced lead times for systems through a consolidated parts bucket strategy.
Distinction between off-the-shelf server hardware and purpose-engineered storage controllers.
Internal bus and memory access optimizations specifically for storage workloads.
The interaction between hardware and software in autonomous operations drives significant performance gains.
[10:00-15:00] Cybersecurity and Ransomware Resiliency
As security becomes a primary objective for storage administrators, the discussion highlights the platform’s new detection and recovery capabilities.
Enhancing cyber resiliency within primary storage rather than relying solely on secondary backup.
Native ransomware detection using anomaly patterns and signature monitoring.
Reducing Recovery Point Objectives (RPO) and Recovery Time Objectives (RTO) through faster detection.
Granular telemetry monitoring at the individual IO level for application security.
Maintaining peak storage performance while running background analysis for security threats.
[15:00-20:00] AIOps and the Shift Toward Automation
The speakers explore the cultural shift in storage administration from Manual “knob-turning” to trusting machine-driven automation and intelligence.
Evolution of storage administrator roles from specialists to infrastructure generalists.
Automating routine tasks like software updates and security patches to reduce admin effort.
Building trust in infrastructure to perform prescriptive remediations automatically.
The impact of cloud-native experiences on user expectations for on-premises management.
Future roadmap for AIOps to provide increasingly actionable intelligence and fewer manual approvals.
[20:00-2500] Enterprise Data Services and Legacy Integration
The segment looks back at the foundational technologies of the past 32-bit and 64-bit file systems while defining modern enterprise requirements.
The architectural heritage of modern file code is rooted in legacy platforms like VNXe.
Challenging the “midrange” label by focusing on enterprise-grade resiliency and data services.
Advanced replication topologies and active-active failover capabilities.
Integration with modern orchestration layers, including Kubernetes through CSI and CSM.
Empowering failover and failback operations at the orchestrated level without manual storage intervention.
[25:00-30:00] Segment 6: Infrastructure Flexibility and Modern Ecosystems
A discussion on how storage must now fit into a complex ecosystem of diverse hypervisors, containers, and multi-cloud environments.
Shifting from evaluating storage in isolation to assessing its fit within the entire infrastructure stack.
Simplifying deployment through pre-built playbooks for Ansible and Terraform.
Managing the complexity of modern environments involving Nutanix, Azure, and Kubernetes.
Prioritizing product roadmaps based on user pain points and market telemetry.
The move toward a faster software release cadence is to stay aligned with evolving tech trends.
[30:00-35:00] Maintenance, Migration, and Life Cycle Management
The podcast concludes with a focus on ease of serviceability and the simplified path for migrating data to the latest generation of hardware.
Serviceability improvements, including intuitive latches and accessible boot drives to minimize downtime.
Zero-impact migration paths through inter-cluster migration tools.
Preserving path management at the cluster level during hardware refreshes.
Economic strategies for utilizing older gear as Tier 2 or Tier 3 storage.
Future-proofing the architecture to support subsequent controller upgrades and technology generations.
The post Podcast #149: Dell PowerStore Elite: Hands-on Deep Dive From the Hopkinton Lab appeared first on StorageReview.com.
39 min
May 21Podcast #148: LinkedIn Live – AI is a Data Problem
StorageReview has published several review pieces and a podcast based on research conducted at Oregon State University. You can find links to those articles at the bottom of this transcript. The articles have generated significant interest not only in research but also across the vendor community. The conversation then expanded to how educators use and track AI in the classroom and during testing, with a focus on how they measure knowledge and learning quality.
The discussion centers on how all this technology is coming together to create real AI outcomes, not just in the scientific research realm, which we’ll get into, but student outcomes at Oregon State, and how these components all come together to make phenomenal solutions.
Due to the overwhelming response to the initial review conducted with Christopher Sullivan, we decided to hold a live LinkedIn event to discuss advances in AI outcomes across both research and education. Our guests included:
Christopher Sullivan – Director of research and academic computing for the College of Earth, Ocean, and Atmospheric Sciences at Oregon State University.
Alan Bumgardner – Director of Associate planning and pathfinding at Celadon Corporation.
Seamus Jones – Director of technical marketing engineering for compute, networking, and sustainability at Dell.
Our participants in this live event were involved in the OSU plankton research, either through direct association (Chris) or in vendor-related roles. If you haven’t read the initial coverage or listened to the follow-up podcast, you’ll find the research both interesting and relevant to today’s environmental focus.
Although the event lasts an hour, there is no marketing or sales pitch, but it is interactive and informative. If you don’t have time to listen to the podcast from start to finish, we’ve broken it down into five-minute segments so you can hop around to the sections that interest you.
00:00–05:00 — Introduction and Context Setting
Brian Beeler opens with a technical, no-sales discussion focused on real-world AI outcomes across research and education at Oregon State.
Panel introductions establish cross-domain expertise spanning academic computing, storage innovation, and enterprise infrastructure.
Oregon State’s mission is outlined: advancing scientific research while improving student engagement through AI.
Chris Sullivan highlights a long-standing involvement that bridges research and academic IT, emphasizing the practical deployment of emerging technologies.
Early framing positions AI as a data-centric challenge rather than purely a compute problem, setting the tone for the session.
05:00–10:00 — Data Scale and AI Workload Realities
Large-scale scientific workloads, such as plankton imaging, generate hundreds of terabytes of data per experiment and require rapid processing.
Training represents a small portion of the total workload; inference dominates at a massive scale, often involving billions of data points.
Temporal relevance drives infrastructure requirements, as delayed insights reduce scientific value.
Continuous feedback loops integrate inference results back into training models, increasing accuracy over time.
Simulation workloads remain critical, feeding data into AI pipelines and further expanding compute demands.
10:00–15:00 — Edge AI and Infrastructure Flexibility
Shift from centralized training clusters to distributed inference at the edge, including deployments on ships and in remote environments.
Flexible platforms, such as GPU-dense servers with high-capacity SSDs, enable multi-environment deployment.
High-density storage allows petabyte-scale datasets to be processed locally, reducing reliance on centralized data centers.
Real-time processing enables adaptive data collection, improving efficiency and reducing operational costs.
Edge inferencing introduces new design considerations for ruggedization, scalability, and modularity.
15:00–20:00 — Enterprise AI Evolution and GPU Utilization
Industry shift from early AI investments focused on training toward widespread inference-driven deployments.
Traditional enterprise applications are becoming AI-aware, leveraging GPUs for performance gains.
Token usage and workflow complexity are increasing significantly as agent-based and multi-step AI processes become more prevalent.
Infrastructure must be designed for bursty, unpredictable workloads rather than steady-state demand.
Planning for scalability and future growth is now essential, even for initial deployments.
20:00–25:00 — Data Management and RAG Challenges
Data accessibility emerges as a primary bottleneck, with large datasets often underutilized due to poor metadata and discoverability.
Retrieval-Augmented Generation requires meaningful metadata, not just file-level attributes.
Scientific and enterprise environments face similar challenges in understanding and monetizing stored data.
Redundant datasets and a lack of visibility highlight inefficiencies in current storage practices.
Effective data curation is positioned as foundational to unlocking AI value.
25:00–30:00 — AI in Education and Secure Data Handling
Oregon State deploys AI-driven student evaluation systems using video-based submissions and automated assessments.
FERPA compliance and data privacy requirements drive on-premises infrastructure decisions.
High-capacity, local storage enables secure handling of large volumes of student-generated content.
Systems must support continuous ingestion, processing, and retention across multiple courses and timeframes.
AI enables scalable student feedback without increasing faculty workload.
30:00–35:00 — Cloud vs On-Prem and Token Economics
Rising cloud costs and token consumption are driving workloads back on-premises, especially for inference.
Tokenomics becomes a key operational consideration, particularly in academic environments with budget constraints.
Hybrid strategies emerge, using on-prem resources for experimentation and cloud for targeted, high-value tasks.
Query optimization and token reduction techniques are becoming critical for cost control.
Data sovereignty and compliance concerns further reinforce on-prem adoption.
35:00–40:00 — Model Optimization and Tiered AI Architectures
Tiered model strategies leverage smaller, specialized models for routine tasks and larger models for complex queries.
CPU and GPU collaboration enables more efficient resource utilization across workloads.
On-prem systems allow iterative experimentation without cost penalties, improving workflow efficiency.
Caching and memory optimization reduce recomputation and improve performance.
Future architectures will increasingly automate model selection and orchestration.
40:00–45:00 — Infrastructure Intelligence and Operational AI
AI is being embedded into infrastructure management, enabling predictive monitoring and automated remediation.
On-prem AI agents monitor campus systems, identifying failures before they impact operations.
Smaller, targeted models are proving effective for operational intelligence use cases.
Hardware advancements, particularly in GPUs, are delivering rapid performance gains with improved efficiency.
Power constraints are shaping system design, prioritizing efficiency over raw scale.
45:00–50:00 — Power, Cooling, and Data Center Design
Power density and cooling are now primary constraints in AI infrastructure deployment.
Liquid cooling enables higher performance, reduced noise, and improved energy efficiency compared to air cooling.
Retrofitting existing data centers with liquid cooling is viable and cost-effective.
Storage efficiency improvements allow more power to be allocated to compute resources.
Cooling innovations, including cold plates and immersion, significantly reduce total energy overhead.
50:00–55:00 — Performance Gains and Practical Deployment
Liquid cooling not only improves efficiency but also enables higher sustained performance through BIOS tuning.
Reduced acoustic impact allows AI systems to be deployed in non-traditional environments.
Universities and enterprises can achieve meaningful ROI through power savings and performance gains.
Cooling strategy selection requires balancing complexity, cost, and operational goals.
Real-world deployments validate the performance benefits of advanced cooling approaches.
55:00–End — Adoption Strategies and Democratizing AI
Lowering barriers to entry is critical, with on-prem sandbox environments enabling risk-free experimentation.
Partnerships with vendors and ISVs are essential for integrating hardware, software, and workflows.
Programs like hardware loaner systems allow organizations to test AI infrastructure before committing.
Pre-validated software stacks simplify deployment and accelerate time to value.
Broad adoption depends on enabling smaller institutions to start small, iterate quickly, and scale as needed.
The following are links to previous articles related to the research conducted by Oregon State University.
AI at Sea: Oregon State Sets Sail to Study the Ocean in Real Time
Podcast #145: Why Edge AI Matters with Oregon State University
How Metrum AI and Oregon State University Are Building the New Standard for Academic Assessment
The post Podcast #148: LinkedIn Live – AI is a Data Problem appeared first on StorageReview.com.
1 hr 3 min
March 03Podcast #147: Inside HPE’s New Model for High-Performance Data Protection
In this week’s podcast, Brian journeyed to Raleigh to visit with HPE’s Chad Bowninger. Chad is Principal Product Manager for Solutions focused on HPE Alletra Storage MP X10000 and the new data protection accelerator node (DPAN).
This is a casual conversation despite the technical topic. Chad even brought in a sofa to make it even more friendly.

The discussion covers everything from the Alletra Storage MP B10000 and its progression to the MP X10000, to data protection and the data protection accelerator node.
The Storagereview team has been working with Chad on a deep dive into the HPE Alletra Storage MP X10000, and will go live on February 26. The review and podcast will give you a solid understanding of the MP X10000 and DPAN.
This podcast is only about 30 minutes long, but we broke it into five-minute segments to give you a chance to jump to topics of interest.
0:00–5:00: Alletra Storage MP X10000 as a Backup Target
Introduction to X10000: Built from the ground up as HPE IP for unstructured data, positioned as the next chapter in the B10000 block platform.
Platform Strategy: Shifting the X10000 from a primary storage role to a high-performance, all-flash backup target.
AI Data Demands: AI pipelines and model training require a massive “workspace” and long-term data retention.
High-Speed Connectivity: Current support for 100GbE with 200GbE certification underway to eliminate network bottlenecks.
5:00–10:00: Data Protection Accelerator (DPA) Architecture
Efficiency over Raw Speed: The DPA node prioritizes effective capacity and operational value over raw IOPS.
Catalyst Technology Integration: Leveraging StoreOnce Catalyst code to achieve deduplication ratios averaging 20:1.
Offloading Heavy Lifting: DPA handles deduplication and encryption, streaming processed data directly to the X10000.
Rapid Recovery Focus: Utilizing flash-to-flash architecture to accelerate rehydration and restore times significantly.
Partner Ecosystem: Designed as a flexible target for existing software like Commvault and Veeam.
10:00–15:00: Linear Scaling and Performance Benchmarks
Modular Growth: Each DPA node manages approximately 2PB, scaling compute and capacity independently.
SLA-Driven Sizing: Adding multiple DPA nodes allows for parallel streaming to meet aggressive RTO/RPO targets.
Ingest Performance: Quoted rates of 1.2PB per hour using a four-accelerator configuration.
Market Differentiation: HPE identifies the DPA as a unique hardware-accelerated approach not currently seen in the market.
15:00–20:00: Portfolio Strategy and Data Value
Enterprise Scale: Targeting multi-petabyte environments in banking, government, and large-scale commercial sectors.
Futureproofing: Disaggregated architecture prevents forklift upgrades by allowing simple drive or node additions.
Immutability and Security: Table-stakes features like immutability are built in to address ransomware concerns.
Multi-Purpose Utility: The X10000 can simultaneously handle backup, AI pipelines, and data lakes.
20:00–28:00: Lab Validation and Business Impact
Real-World Testing: Lab configurations utilize 30-host VMware clusters to validate performance at scale.
Network Evolution: Bridging the gap between high-speed AI fabrics and traditionally slower backup networks.
The “Unspoken Reality”: Addressing the common gap between official SLAs and actual, slower recovery capabilities.
Economic Justification: Linking rapid restore performance directly to minimizing revenue loss during downtime.
Predictable Forecasting: Linear scaling simplifies budget and capacity planning for growing organizations.
The post Podcast #147: Inside HPE’s New Model for High-Performance Data Protection appeared first on StorageReview.com.
29 min
December 29, 2025Podcast #145: Why Edge AI Matters with Oregon State University
If you follow StorageReview regularly, you might have seen the piece on the OSU Ocean Study in real time. We spoke with Chris Sullivan, Director of Research and Academic Computing at Oregon State University, to better understand how technology helps researchers study ocean life and how that impacts the global environment.
Brian met up with Chris at SC25 to discuss the topic further. It’s rare for us to write about how technology is used in real-world situations such as OSU’s plankton research, so this feature and the podcast gave us that opportunity.
As Brian and Chris discuss the research being conducted at OSU, there’s genuine excitement about how technology is advancing to support ongoing oceanic studies. Chris has been with Oregon State University for over 24 years and has served as director of biocomputing for the past three years.
This live podcast showcases Chris’s excitement for science and technology. He called SC25 “candy land.”

This is a lively and honest discussion (with some humor) that provides genuine insights into how research uses technology to deliver real-time results.
It’s a short podcast worth listening to, but if you’re pressed for time, we’ve provided a transcript that lets you jump to the most relevant sections.
0:00 – 5:00
Turning a Research Vessel into a Floating Supercomputer
Chris Sullivan from Oregon State University explains how new tech is fundamentally changing how ocean science is done, especially massive plankton studies run from ships that are basically small, floating data centers.
Technology isn’t just “cool gear”; it directly changes what science is possible and the extent of bias in results.
Scientific bias: limited or skewed samples leading to inaccurate conclusions; the cure is more data points across more locations.
Goal: Collect huge, diverse datasets so statistical analyses aren’t distorted by sampling just one spot or one set of conditions.
Traditional HPC setups forced a trade-off: fast local NVMe/SSDs with limited capacity vs. large but slower tier-one storage systems.
For large experiments, Chris had to depend on tier-one storage across the cluster because on-node (tier-zero) storage was too small to hold massive datasets.
Big, shared-storage “super pods” became necessary to keep GPUs fed, but they’re expensive and not realistic to deploy in the field or at sea.
5:00 – 10:00
Ocean Robots, Plankton, and a Million-Dollar Cruise
This section dives into what actually happens on the ship—how they tow instruments, capture staggering amounts of image data, and why semi–real-time processing is critical when ship time costs about $1M per 10‑day cruise.
The new NSF-backed research vessel is built like a “mini Star Trek Enterprise,” packed with tech and about 200 miles of cabling on a 200‑foot boat.
Previously, Chris had to roll his own compute hardware onto the ship, string fiber down hallways, and operate without a proper on-board data center.
The plankton system: a towed device moves up and down through the water at 5 knots, pulling ~162 liters of water per second and generating ~10 GB of image data per minute.
A single 10‑day plankton experiment: ~100 TB of raw data; after processing, segmentation, and classification, that expands to ~300–400 TB of usable data.
They must process this in semi–real time; otherwise, a $1M, 10‑day cruise could return with bad or useless data and no opportunity to redirect the ship.
Plankton are mostly at the mercy of currents (not strong swimmers), yet they’re crucial: ~50% of Earth’s oxygen, ~25% of the carbon sink, ~17% of global food per capita, and the base of the marine food web.
10:00 – 15:00
Edge AI at Sea and in the Forest
The focus shifts from the ship to the broader “edge” problem: how to push compute closer to where data is created—on boats, in ranger stations, and across forests filled with autonomous sensors.
New high‑capacity SSD-based systems let them build a true tier zero on a single device: store hundreds of TB of data, process in semi‑real time, and never move it again.
Earlier tests with smaller NVMe drives required constant offloading to make room, which burned precious IO bandwidth and slowed the science.
There is essentially no dedicated IT staff on the ship; one person runs the ship systems and stays away from experiments, so grad students and postdocs must manage hardware and pipelines themselves.
This drives a design requirement: simple, robust pipelines and self-contained devices—no fragile web of arrays, mounts, and volumes.
At the forest edge (e.g., the Owl project), they run ~5,000 autonomous recording units and can identify ~130 bird species by sound alone.
The pain point: humans still have to go collect SD cards; by the time the data returns, much of the temporal value is gone. Edge processing aims to extract insights immediately at the point of capture.
15:00 – 20:00
Matching the Right GPUs and Storage to the Job
Here, they unpack GPU strategy, Omniverse-based training, and why large-capacity flash is as important as raw compute for AI and scientific visualization.
Chris stresses using the proper GPU for the job: large, expensive accelerators (e.g., H100/Hopper, “grasshoppers”) should focus on training, while more cost-effective cards like RTX 6000s handle inference at scale.
Historically, there was confusion: too many overlapping GPU SKUs, unclear “swim lanes,” and a lot of money wasted on overkill hardware for simple inference.
RTX 6000s give them a dual role: high-performance inference plus graphics and rendering capabilities that data-center-only GPUs lack.
They’re building an Omniverse digital twin of the ship to:
Train students and postdocs virtually while the ship is at sea.
Plan physical layouts and equipment loading in tight, moving spaces.
Avoid wasting expensive ship time on basic training and trial‑and‑error setups.
On a moving ship, anything with moving parts is a risk; solid-state storage is preferred for reliability and resilience.
High-capacity SSDs in a small power envelope free up power budget for more GPUs, reduce the need for separate arrays, and shrink the total hardware footprint.
20:00 – 25:00
From Filing Cabinets to Fiber-Connected Oceans
Zooming out, how cheap storage unlocked modern AI, how ocean observatories stream data ashore, and why network and storage together are enabling new kinds of science.
Chris argues AI’s real unlock was not new math—most models date back to the 1950s–70s—but the cheap hard drive of the 1990s, which moved data from filing cabinets onto searchable, computable media.
In the “filing cabinet era,” they had tons of data but were “information poor”—no search, no redundancy, and no scalable processing.
There’s always been an arms race between processing speed and data availability; large-capacity SSDs now tip the balance again by feeding modern GPUs with far more data.
Oregon State is the cyberinfrastructure lead for the Ocean Observatories Initiative (OOI), a billion‑dollar NSF/WHOI project that uses gliders, buoys, and seafloor mounts to stream ocean data.
They hold ~9 PB of OOI data, with a fiber backbone upgraded from 200 Gbps to 400 Gbps via Link Oregon, enabling real-time data transfer from the ocean to the data center.
Even with Starlink on the ship and on land projects (like the owl work), satellite bandwidth and cost limit what can be sent, so they’re testing LoRaWAN and other mesh networks to push data back to a single, Starlink‑connected ranger station.
25:00 – 27:52
Lessons for Enterprise: Your Data Center Is the Edge Now
The conversation closes by translating these research lessons to enterprises and highlighting the hardware Chris would “steal” from the show floor.
Core message to enterprises: if you’re not processing at the edge, you’re burning time and labor hauling raw data back and forth. And missing real-time opportunities.
Edge AI and agents should preprocess, filter, and analyze locally to redirect human effort from logistics to higher‑value tasks and decision-making.
Real‑world analogs:
Retail uses edge analytics in stores to track traffic, make recommendations, and deliver dynamic experiences, much like forestry uses edge inference to protect endangered species while enabling logging.
Fisheries want “smart traps” and buoys so they only send out boats when there are actually crabs in the pots.
For Chris, the ideal pattern is on-premises data centers that become edge facilities integrated with the cloud for large training runs and longer‑term workloads.
Favorite hardware from the show: Dell’s PowerEdge 7745; he wants more of them for VR labs and field deployments because they strike the right balance of power and footprint, flexibility, and capabilities (rendering + inference + edge work).
He sees systems like the 7745 as the template: flexible enough to live in a data center or in harsh field conditions, supporting both traditional HPC and advanced AI/visualization at the edge.
The post Podcast #145: Why Edge AI Matters with Oregon State University appeared first on StorageReview.com.
28 min
November 24, 2025Podcast #144: 300 GB/s in One Dell PowerEdge R7725xd
At the SC25 event in St. Louis, Brian caught up with his long-time friend Mark Klarzynski, founder and acting CTO at PEAK:AIO. If you have been following StorageReview, then you may be familiar with our coverage of PEAK:AIO, from regular news coverage to medical AI to animal conservation.

As Mark explains, he has “spent a lifetime in storage, and he has known Brian for a very long, long time.” It shows with a very candid, easy conversation about hardware and software, and the evolution of both.
In this really short Podcast, Brian and Mark talk about software development, Dell’s PowerEdge, and GPU servers. Mark is excited about advances in hardware design, especially Dell’s PowerEdge R7725xd, powered by AMD EPYC CPUs.
This short podcast covers a lot of territory, and it is worth listening to in one sitting. However, if you can’t spare 20 minutes, we’ve broken it down into five-minute sections so you can hop around.
0:00 – 5:00
PEAK:AIO software unlocks the Dell 7725XD’s extreme performance.
Modern hardware (CPUs, NVMe, networks) is finally good enough that software can “get out of the way.”
PEAK:AIO removed decades of legacy storage features to get closer to bare metal.
Their surprise: the Dell box delivered far higher performance than expected; they thought the tests were wrong.
Philosophy: Making things simple and stable is more complicated than just making them fast.
5:00 – 10:00
PEAK:AIO chose to remain independent rather than become a feature within a prominent vendor’s product line.
They work directly with research teams (e.g., brain cancer models), which is more rewarding than just speeding up databases.
Early focus: compressing what previously required 4–10 storage nodes into a single powerful server.
Traditional parallel file systems were built for 10GbE + HDD; NVMe and massive PCIe bandwidth are now required.
With Los Alamos, they’re building an open, modern NFS fabric—no proprietary lock‑in.
10:00 – 15:00
Hardware “wish list”: more PCIe lanes and PCIe Gen 6 to unlock even more bandwidth.
Vision: GPU servers that are a model box, inference box, and storage node all in one, yet part of a larger namespace.
Start small on a GPU server; when storage runs out, add a 7725XD and let PEAK:AIO automatically expand and rebalance.
Customers want to scale storage only when AI projects prove value, not buy a giant array up front.
Economics matter: you can’t justify a $1M storage array behind a $300k GPU server for hospitals and similar budgets.
15:00 – 20:00
Strategy: Use commodity servers plus PEAK:AIO software to replace expensive, specialized storage arrays.
OEMs like Dell handle fundamental hardware issues; PEAK:AIO focuses on software, stability, and performance.
Built-in monitoring and predictive intelligence; years in the field with no downtime from their platform.
Joint project results: ~70M IOPS in a single node and hundreds of GB/s throughput—record numbers for them.
Next step: put models inside the storage node (with spare CPU/GPU capacity) so inference happens in-box, not over the network.
The post Podcast #144: 300 GB/s in One Dell PowerEdge R7725xd appeared first on StorageReview.com.
20 min
November 16, 2025Podcast #143: KIOXIA 245TB SSDs Are Here!
KIOXIA’s Maulik Sompura joins Brian for an in-depth, informative discussion about all things flash and other industry events. The topic is timely and relevant, given the massive expansion of AI and modern workloads.

Maulik Sompura is the Senior Staff Director—Product Planning and Management at KIOXIA, with over 13 years of experience in NAND, memory, and SSDs, including the last six years at KIOXIA. Additionally, his career includes stints with Intel and Toshiba.
The conversation extends beyond storage and memory, transitioning to liquid cooling, hyperscalers, and form factors. They also discuss open-source software, new fabs, concerns about scaling, and fears of another “Covid-boom-and-bust” event.
It is an informative conversation that will prompt you to think about something in the long term. We’ve broken the podcast into five-minute segments, so you can jump to the section that’s most relevant to you and your environment. However, it is well worth the time to watch this in its entirety.
Brian and Maulik kick off the discussion with how radically SSDs and flash storage have evolved, especially under the pressure of AI and hyperscale data centers.
0:00–5:00: Hyperscalers Quietly Running the Show
SSD portfolios have exploded in complexity: many form factors, endurance levels, and performance tiers.
The AI boom has moved storage right behind GPUs/HBM as a first-class priority.
Hyperscalers (Meta, Google, Amazon, etc.) dictate detailed drive requirements, and those custom designs ultimately shape what enterprises get.
OCP specs help unify and simplify the chaos by giving everyone a common target on top of NVMe.
The industry is actively trying to shrink the number of SKUs and converge on modern form factors like E3, even as legacy form factors linger.
5:00–10:00: E3, E2, and Petabyte-Class Drives
E3 is becoming the preferred enterprise form factor for Gen6+ due to its better power and thermal characteristics.
2.5″ drives will stick around for years due to long product lifecycles, but new designs are moving to EDSFF.
E2 is a larger, capacity-first SSD form factor, seen as a true hard drive replacement at extreme capacities.
With advanced NAND stacking, a single E2 drive could eventually reach 1 PB.
Hyperscalers mitigate the risks of giant drives using sharding, erasure coding, and predictive analytics; rebuild times for huge SSDs are still much better than for large HDDs.
10:00–15:00: Liquid Cooling Comes for SSDs Too
AI training systems are driving widespread adoption of liquid cooling, and that now extends down to storage.
SSDs in E3 form factors are being redesigned for direct contact with cold plates (surface flatness, materials, and connector details).
Liquid-cooled Gen5 SSDs will pave the way for Gen6, where power and heat go even higher.
Enterprises are cautious about bringing water into the data center, but rising power costs and GPU density will make liquid cooling harder to avoid.
15:00–20:00: AI Storage: Capacity, Performance, and Fabric Stress
AI pushes storage along two axes: massive capacity and very high performance.
KIOXIA’s 245 TB-class drive (with 32‑high QLC stacks) targets data lakes and large repositories, not just generic workloads.
Inside a modern box, storage can push 250–280 GB/s; the bottleneck increasingly becomes the network fabric, not the drives.
Hyperscalers are racing to 400/800 GbE and smarter NICs/DPUs, while mid-market customers struggle to keep up with that level of fabric investment.
Traditional databases and Fibre Channel remain, but vector databases and AI-centric data stores are joining the mix.
20:00–25:00: Smarter Software and a True Flash Super‑Cycle
KIOXIA isn’t just shipping drives, it’s investing in software (e.g., ISAAC) to reduce HBM/DRAM needs while preserving AI performance.
The strategy is holistic: improve cost and efficiency at the system level, not just through faster NAND.
New fabs like Kitakami (K2) are ramping up, but vendors are cautious after Covid-era boom-and-bust cycles that cost the industry ~$30B.
Scaling NAND (more layers, finer lithography) is getting more expensive and faster-paced, so that supply will grow, but in a controlled way.
Many see this as a multi-year “super‑cycle” driven by AI and data growth, not just a normal 1–2-year flash-upturn.
25:00–End: AI Everywhere, Flash vs. HDD, and Learning from HPC
Data keeps growing and is rarely deleted; AI is moving to wherever that data lives, including the edge, backups, and secondary tiers.
High-capacity SSDs can deliver ~100 PB per rack, versus multiple racks of HDDs for the same capacity, with far better performance and efficiency.
HDD performance doesn’t scale with capacity; SSDs deliver orders-of-magnitude better IOPS/GB and throughput, improving TCO and favoring flash at scale.
Shows like OCP and Supercomputing (SC) are where the future is visible first: wide racks, 800V busbars, exotic cooling, and power designs.
Hyperscaler and HPC innovations are steadily shaping what “normal” enterprise servers and data centers will look like, with resiliency and efficiency.
The post Podcast #143: KIOXIA 245TB SSDs Are Here! appeared first on StorageReview.com.
40 min
November 12, 2025Podcast #142: What’s Next For Storage in the AI Era?
Brian and crew recently landed in Montreal to tour the Hypertec lab facility and witness some interesting immersion-cooling trends with the leader in the field.
Following the lab tour and product demos, Brian sits down with Scott Shadley, Director of Leadership Narrative at Solidigm. Scott and Brian have a lively, candid discussion as they trace the evolution of cooling, flash storage, and breaking away from legacy hard-drive form factors.
However, it’s not all about immersion technology. Brian and Scott discuss signal integrity challenges in PCIe Gen5/6 and the trade-off between performance and reliability. Of course, there can’t be a podcast without an AI segment, so we cover topics such as AI-driven data growth, RAID resiliency, and large-drive “blast radius” management.

Scott discusses Solidigm’s focus on developing products that customers want, rather than guessing what the industry needs. This only works when vendors work closely with customers to understand their pain points. Scott gets into the collaboration with NVIDIA. It is easy to hear the excitement in Scott’s voice. Of course, because it is a massive hardware environment, Brian has a hard time wiping the smile from his face.
This podcast covers everything from early Fibre Channel drives to flash and NVMe, and also explores edge computing applications, AI inference offload, and immersion-qualified SSD developments.
This pod is a must for anyone interested in the evolution of immersion cooling and flash storage, as well as why a strong partner ecosystem is crucial for customer success. If you have an extra 35 minutes, give this a listen or watch it from end to end; it’s worth the time. If you’d like to explore specific topics, we have broken it down into five-minute segments.
0:00 – 5:00 | Immersion Cooling and Industry Trends
Highlights:
Opening from Montreal with Solidigm and Hypertec.
Discussion of immersion cooling, liquid vs. air cooling, and touring Hypertec’s facility.
Mention of OCP and how data center engineers are “becoming plumbers” due to widespread liquid‑cooled infrastructure.
Solidigm’s collaboration with partners on direct‑to‑chip liquid cooling solutions is co‑designed with NVIDIA.
Philosophy of open innovation: Solidigm contributes IP to industry standards via SNIA to ensure interoperability.
Emphasis on customer‑driven innovation over chasing the next PCIe or interface generation.
5:00 – 10:00 | Evolution of Flash Storage and Form Factors
Highlights:
Reflecting on early enterprise flash (Intel X25, SAS/Fibre Channel SSDs).
Transition from legacy hard drives to SSDs and the “breaking the box” moment with NVMe.
Emergence of new form factors (EDSFF → E1.S, E1.L, E3).
Discussion about how standardization is evolving beyond traditional 2.5” and 3.5” drive designs.
Design challenges with immersion environments and high‑density flash platforms.
Importance of thermal stability: consistent environments improve endurance and reliability.
Cooling innovations directly tie to longer SSD lifespans and greater efficiency.
10:00 – 15:00 | PCIe Generations, Signal Integrity, and System Design
Highlights:
Talk about connector design challenges in next‑gen cold‑plate and PCIe Gen 5 and 6 environments.
Signal integrity issues grow from PCIe 3→5→6: Every generation increases susceptibility to noise.
Analogy: PCIe evolution from 3G to 5G: more nodes, shorter distance, tighter tolerances.
Solidigm’s view: optimize system architecture and density instead of chasing raw drive speed.
Many enterprise customers remain on Gen4 because it’s “fast enough” for distributed workloads.
Shift from pure performance toward holistic throughput, reliability, and efficiency.
15:00 – 20:00 | Reliability, Capacity, and Customer Trust
Highlights:
Performance is balanced with quality, longevity, and consistency.
SSDs reaching “reliability of a DIMM.” Many customers stop swapping drives; they fail “eloquently.”
Customer satisfaction and trust are top priorities; Solidigm scores highly in vendor ratings.
Rise of ultra‑dense drives like 122TB models (D5‑P5336 family) and discussion of adoption.
Larger capacities are no longer halo products.
Manufacturing challenge: Smaller drives are harder to produce as NAND density increases.
Drives sized for total cost of ownership (TCO) optimization; approach to “replace spinning disks.”
20:00 – 25:00 | Data Growth, AI, and Blast Radius Concerns
Highlights:
Industry shift from deleting data to keeping everything, fueled by AI training and inference needs.
“Data sovereignty” trend: Regional training data sets lead to more localized capacity requirements.
Growing demand for massive local storage to adapt global models like GPT.
Discussion of “blast radius” and the risk of losing a 100 TB drive.
Hardware RAID for NVMe is back; GPU‑accelerated parity rebuilds are gaining traction.
SSD rebuilds are fast. Downtime rather than permanent data loss is the genuine concern.
Many operators now design for failure‑in‑place architectures instead of manual swaps.
25:00 – 30:00 | Future Directions, AI Edge, and Innovation Opportunities
Highlights:
Reflection on industry-wide stock shortages driven by the AI boom and NAND supply lags.
Unique opportunity to rethink architectures and possible new form factors or integrated flash + networking.
Edge innovation: NVIDIA’s Spark and Jetson products blur the boundaries between data centers and desktops.
Exploration of NVMe offload for inference and retrieval-augmented generation (RAG).
Hope for a future where storage innovations take center stage at GTC alongside compute and DRAM.
Emphasis on cross‑industry collaboration. SNIA’s new “Storage.AI” working groups integrate storage, networking, and AI alignment efforts.
30:00 – 34:00 | Closing Thoughts and On‑Site Testing
Highlights:
Returning to Hypertec’s immersion servers for real‑world tests with Solidigm drives.
Discussion of thermal stability advantages: consistent temperatures equal no throttling.
Solidigm is working toward fully qualified SSDs for hydrocarbon‑based immersion liquids.
Closing conversation: Mutual excitement about hands‑on experimentation, performance consistency, and hardware innovation.
Wrap‑up and goodbyes from Montreal.
The post Podcast #142: What’s Next For Storage in the AI Era? appeared first on StorageReview.com.
35 min

About StorageReview.com

From the publisher's feed

StorageReview.com is a leading provider of news and reviews throughout the entire IT stack - from the datacenter to the edge, and all points in between.

Share StorageReview.com

Sign up to save your podcasts

StorageReview.com

StorageReview.com episodes

About StorageReview.com