Hammerspace has added its data orchestration software in Arrow Electronics-supplied Dell server boxes. Customers can buy subscriptions to these appliances to get their scattered data estates orchestrated and organized for accelerated access.
Hammerspace’s Global Data Platform (GDP) includes parallel NFS-based file system software to manage data in globally distributed locations, both in file (NFS) and object storage, using SSDs, disk drives, public cloud services (AWS, Google Cloud, Azure, and Seagate Lyve Cloud), and tape libraries. The data can be located in a global namespace, orchestrated, placed, and accessed as if it were local. Hammerspace is working with Dell servers and Arrow Electronics to have reference architecture-defined appliances made available to customers for faster deployment of systems. Customers or resellers can order pre-configured bundles based on the most common deployment scenarios, such as size, number of locations, and performance requirements.
Molly Presley, Hammerspace SVP of Global Marketing, said: “Organizations today need streamlined data strategies that empower their analytics teams to pursue AI, data analytics, and deep learning, all while ensuring robust data availability and security. Achieving this balance requires the high performance of parallel file systems for GPU-intensive compute workflows, the enterprise-grade features of NAS to meet industry standards, and the unifying power of a global namespace to seamlessly integrate data across locations.
“Our new series of appliances marks a critical leap forward in accelerating the deployment of our Global Data Platform software, delivering the capabilities our customers need to stay ahead.”
Hammerspace’s GDP has separate Anvil metadata server and data storage components, and the appliance reference architecture makes this clear. There are a series of defined appliance models:
The metadata and data servers are scaled independently, and there are two types of data servers, Data Services (DSX) and Linux storage server nodes. DSX nodes scale out client access and performance, providing multi-platform data services including multi-protocol access: NFS v3, SMB and S3. The Linux storage server nodes are for direct parallel NFS v4.2 access, scaling performance and capacity, or just to scale capacity:
There are two types of appliance; those optimized for sheer throughput performance or focused on capacity:
The data servers are either a 1RU x 10-bay drive server model or a 2RU x 24-bay model. A variety of NVMe SSDs are available, both mixed use (1.6-6.4 TB) and read-intensive (1.92-61.44 TB).
Each of the metadata and data servers can be configured with 25, 100, 200 or 400 GbitE network cards.
Check out a Hammerspace Dell server webpage here and the seven-page reference architecture document for more details. Hammerspace software can also run on any of the popular NAS, block, object, and cloud storage platforms.
IBM has combinined data lakehouse and parallel file system features to provide a scalable, machine-learning-based AI processing and analytics data storage platform, using its watsonx.data and Storage Scale products.
This delivers, IBM claims, extreme AI performance with GDS (GPU Direct Storage) and Nvidia for training generative AI models faster. There is multi-protocol support enabling simpler workflows, a unified data platform for analytics and AI, and the system supports retrieval-augmented generation (RAG) using a customer’s proprietary data.
Big Blue’s watsonx.data is a data lakehouse. It combines the features of a data lake – based on a scale-out architecture using commodity servers and capable of storing and processing large volumes of structured and unstructured data – with the performance of a data warehouse. It supports the Apache Iceberg open table format enabling different processing engines to access the same data at the same time.
Storage Scale is a parallel and scale-out file system originally called GPFS. It is being used as a storage layer underneath watson.x.data, providing object storage facilities underneath a file access overlay. V5.2.1 Storage Scale has a non-containerized, High Performance S3 protocol service providing this.
How it fits together
An IBM diagram, with added B&F yellow box elements, lays out the software componentry:
There are separate, disaggregated, compute and storage layers. For compute, a Red Hat OpenShift container cluster base is used by watsonx.data applications, which include Presto and Spark. Presto provides datalake analytics using a distributed SQL query engine while Spark is an in-memory big data processing and analytics resource.
There is also a Hive Metastore providing a shared metadata service and a Milvus vector database service in this layer. Milvus is used to enable RAG by accessing a customer’s potentially large datasets residing on Storage Scale.
There are three main elements in the separate storage infrastructure: Storage Scale filesystem clusters holding the data; Active File Management (AFM) for storage abstraction and acceleration; and S3 data access protocol service for high performance object access.
The S3 service exposes object storage buckets to watsonx.data for attachment to a query engine such as Presto or Spark. S3 objects are mapped to files and buckets are mapped to directories within Storage Scale and vice versa.
The S3 buckets can be local to the storage layer or cached (and thus accelerated) by Storage Scale from external object stores which may be globally dispersed across various clouds, datacenters and locations. In either case, multiple Spark and Presto engine instances connect to the Storage Scale layer using the S3 protocol to access the buckets.
AFM has local caching and enables sharing of data across clusters, virtualizing remote S3 buckets at fileset level. It implements a global namespace across Storage Scale clusters and can include NFS data sources in this namespace as well.The remote S3 buckets appear as local buckets under the Storage Scale file system, under a common storage namespace. This eliminates the need for data copies.
The virtualizing of remote S3 buckets relies on Storage Scale High Performance S3, which is based on NooBaa open-source software. This is object storage software using X86 servers and storage, presented as an S3-like cloud service. Noobaa was acquired by Red Hat in 2018, and abstracts storage infrastructure across hybrid multi-cloud environments. It also provides data storage service management. Red Hat made it part of its OpenShift Data Foundation (ODF) product set. IBM bought Red Hat in 2019 and added ODF to its then Spectrum Fusion offering – now Storage Fusion – alongside the existing containerized version of Spectrum Scale and Spectrum Protect data protection.
Now NooBaa is a customizable and dynamic data gateway for objects, providing data services such as caching, tiering, mirroring, dedupe, encryption, and compression, over any storage resource including S3, GCS, Azure Blob, filesystems, etc.
Storage Scale’s High Performance Object S3 service is optimized for multi-protocol data access. It replaces earlier Swift-based Object S3 and Containerized S3 service implementations in Storage Scale. A Cluster Export Services (CES) facility within Storage Scale manages high-availability through CES nodes.
Multiple tiers
IBM says there can be multiple performance tiers for Storage Scale storage, to optimize costs and performance. There can be a high-performance tier for hot data, along with a cost-effective tier or even tape for long-term storage and archival, together with automated policy driven placement across tiers – making the tiering seamless and transparent to applications.
This combined watsonx.data and Storage Scale system provides a unified but disaggregated compute and storage platform on which to run AI applications for both training and inference. Customers may well value this as IBM is acting as a single source for the required software. We have covered other AI data platform approaches from Dell, HPE, Lenovo, NetApp, MinIO and Pure, with VAST Data preparing its own Data Engine offering.
The watsonx.data and Storage Scale AI bundle is described in an IBM Redbook which “showcases how IBM watsonx.data applications can benefit from the enterprise storage features and functions offered by IBM Storage Scale.”
You can find out more about the latest watsonx.data v2.0.2 from release notes here.
NetApp has announced new capabilities that support VMware Cloud Foundation (VCF) deployments, with mutual customers being promised they will be able to “right-size” their IT environments to run VMware workloads efficiently at the right scale.
To help customers reduce costs, simplify operations, and increase flexibility, NetApp claims they will be able to use NetApp ONTAP software for all their VCF storage requirements, including standard and consolidated architectures.
The latest release of ONTAP Tools for VMware (OTV) will support SnapMirror active sync to provide symmetric active-active data replication capabilities for NetApp storage systems running VMware workloads. SnapMirror active sync allows customers to operate more efficiently by offloading data protection from their virtualized compute and improving data availability.
In addition, with new capabilities for Azure VMware Solution (AVS), customers that are extending or migrating their vSphere workloads to the cloud can now leverage Spot Eco by NetApp with AVS reserved instances to get the most value out of their deployments. Using Spot Eco to manage AVS reserved instances, while also using Azure NetApp Files to offload data storage, can reduce compute costs “significantly,” says NetApp.
Also, NetApp is introducing Cloud Insights VM Optimization, which is a set of tools designed to reduce costs by increasing VM density, run storage at the best price-to-performance ratio for company environments, and monitor the entire environment to ensure availability, performance, and adherence to configuration best practices across the full stack.
To help customers optimize the compute, memory and storage resources of their VMware environments, NetApp is offering customers a free 30-day trial of Cloud Insights, to help them “cost-effectively migrate” to newly introduced VMware software subscriptions brought in by VMware owner Broadcom.
Jonsi Stefansson.
“NetApp and Broadcom are working together to take the uncertainty out of hybrid cloud environments,” said Jonsi Stefansson, SVP and chief technology officer at NetApp. “More than 20,000 customers rely on NetApp to support their VMware workloads.”
“As organizations modernize infrastructure with VMware Cloud Foundation, they want to know the services they rely on will continue to work seamlessly and deliver the value they have come to expect,” added Paul Turner, VP of products, VCF division at Broadcom. “Having NetApp as a close collaborator helps mutual customers deploy innovative data and storage services on top of their private cloud platform.”
Earlier this week, NetApp reported a good first quarter, with sales up 8 percent to $1.54 billion, and a 66 percent increase in profits to $248 million.
Rubrik is joining forces with Cisco to enhance organisations’ cyber resilience, with Rubrik integrating with the Cisco XDR (extended detection and response) platform, and joining its SolutionsPlus program.
The integration will make it easier for enterprises to quickly identify targeted attacks on data, and deliver a secure recovery, the partners said.
The partnership will deliver combined threat intelligence and data context to prioritise and respond effectively to threats targeting critical data, as well as streamlined security and IT operations to help repair after attacks.
Automated, multi-site scaling and lifecycle management technology from Cisco Intersight will also enable the deployment and tuned configuration of the Rubrik Data Management Platform on Cisco UCS (Unified Computing Systems) server infrastructure at a global scale.
Rubrik will join Cisco’s SolutionsPlus program as well, with all of Rubrik’s products being added to Cisco’s global price list. As part of SolutionsPlus, Cisco’s direct and channel sales network can now offer Rubrik Security Cloud with Cisco UCS and Cisco XDR systems.
Rubrik’s Data Security Posture Management (DSPM) technology is also on the price list, allowing joint customers to proactively reduce the risk of data exposure and exfiltration across on-premises, cloud, and SaaS environments, through knowing where all the data is and how to secure it, we are told.
Ghazal Asif Farhadi.
“As the volume of data continues to grow and exacerbate visibility challenges, organizations must find ways to manage and protect their constantly expanding data,” said Ghazal Asif Farhadi, VP of global channels, alliances and inside sales at Rubrik. “Therefore, Rubrik and Cisco have joined forces to enhance organizations’ cyber resilience and protect mission-critical data against cyber attacks.”
“Complex, siloed systems create gaps in visibility that hinder operations and slow down organizations’ ability to detect and respond to threats,” added Jeremy Foster, senior vice president and general manager of Cisco Compute. “We are now delivering threat detection and data context in one place, enabling security teams to effectively prioritize critical attacks and respond to them quickly.”
Jim Finn, sales VP, cyber at Presidio, said: “As a major partner to Cisco and Rubrik, we believe their collaboration will help our customers proactively advance and simplify their data security.”
Earlier this month, Rubrik extended its SaaS app data protection to Salesforce Core platform data.
A massive rise in server sales led Dell’s second fiscal 2025 quarter revenues to a 9 percent year-on-year rise to $25 billion, with an $841 million profit, more than Nutanix’ entire $524.6 million quarterly revenue, but there was a slowdown in HCI sales.
Dell said there was: “Exceptional AI-optimized server demand and improved traditional server demand growth,” in the quarter ended August 2, 2024.
Jeff Clarke.
Vice-chairman and COO Jeff Clarke stated: “We executed well in Q2, and I’m really proud of our team and our performance.” The revenue number was “another record for our servers and networking business.”
Quarterly financial summary
Gross margin: 21.2 percent vs 23.5 percent a year ago
Operating cash flow: $1.3 billion
Free cash flow: $1.28 billion vs $3.1 billion last year
Cash, cash equivalents, and restricted cash: $6.0 billion vs $7.9 billion last year
Diluted earnings per share: $1.17, up 86 percent y/y
The $1.7 billion downturn in cash and investments was the result of Dell making capital returns of $1 billion; repurchasing 5.5 million shares of stock at an average price of $130.03 plus paying a $0.45 per share dividend, and net debt paydown of $1 billion during the quarter.
Dell has resumed a growth trajectory after 6 quarters of declining revenue.
The two main Dell business units, the Infrastructure Solutions Group (ISG) and Client Solutions Group (CSG), had differing results with ISG (servers, storage and networking) revenues rising 38 percent Y/Y to a record $11.6 billion, while CSG (PCs, etc.) revenues went down 4 percent to $12.4 billion. In the CSG unit, commercial client revenue was flat at $10.6 billion, and consumer revenue was 22 percent down at $1.9 billion. ISG is now in a position to overtake CSG revenues after years of CSG being far ahead.
ISG revenues could overtake CSG revenues if the AI server sales boost continues.
The ISG revenue driver was and is server sales, with a remarkable 80 percent rise in the combined server and networking revenue number to a record $7.7 billion. Storage revenue was $4.0 billion, down 5 percent.
Note the sudden rise in ISG’s server and networking revenues this quarter.
Clarke’s prepared remarks expanded on this: “Our AI momentum accelerated in Q2, and we’ve seen an increase in the number of enterprise customers buying AI solutions each quarter. AI-optimized server demand was $3.2 billion, up 23% sequentially, and $5.8 billion year to date. Backlog was $3.8 billion, and our pipeline has grown to several multiples of our backlog.” This hints that next quarter’s server revenues should be impressive as well.
There was demand growth in storage, with Clarke saying in the earnings call: “Dell IP core storage demand, including PowerMax, PowerScale, PowerStore and PowerProtect/Data Domain grew double-digits in the quarter, a positive sign as we move into the second-half of the year.”
CFO Yvonne McGill said storage did well under the covers: ”We’re pleased with the sequential improvement in storage profitability. We gained scale with 6 percent sequential revenue growth. We were more disciplined in our pricing. We had a higher mix of Dell IP storage solutions and a better geographic mix with more North America activity.”
She also identified the storage weak point: ”This strong storage performance was offset by headwinds in the partner IP portion of our HCI (hyper-converged infrastructure) portfolio.”
Earnings call analyst Tony Sacconaghi suggested “that hyperconverged is either an enormous part of your business or was down by a huge amount.” Clarke confirmed that: “obviously, your conclusion is correct. The partner IP business was down quarter-over-quarter.”
Dell’s Investor Relations head Rob Williams added: “We haven’t broken it out specifically, but it [HCI] is obviously a large part of the business. And data protection would be the other area where there was a bit of decline year-over-year. So those two offset the increases in the other four businesses we mentioned.” The other four being PowerMax, PowerScale, PowerStore and PowerProtect/Data Domain.
Dell supplies VxRail HCI running vSphere and also PowerFlex HCI systems. There are well known VMware market issues following Broadcom’s acquisition of VMware. Dell signed an HCI partnership deal with Nutanix in the quarter, looking to grow its HCI business.
Back to AI
AI server sales dominated the quarter, with Clarke saying: ”Our unique capability to deliver leading-edge air and liquid cooled AI servers, networking and storage tuned and optimized for maximum performance at the node and rack level combined with leading ecosystem partners and world-class services and support continues to resonate with customers.”
There’s more to come: “Enterprise remains a significant opportunity for us as many are still in the early stages of AI adoption. We are also excited about the emerging sovereign AI opportunity, which plays to our strengths given our position with governments around the world.”
And: ”Most exciting, our AI server pipeline expanded across both Tier-2 CSPs and enterprise customers again in Q2, and now has grown to several multiples of our backlog. As we begin the second-half of the year, we have optimized our sales coverage to better focus on AI opportunities across CSPs, and both large and small customer segments and geographies.”
The outlook is excellent: “Our AI opportunity with Tier-2 CSPs, enterprise and emerging sovereign customers is immense. Our view is supported by an AI hardware and services TAM of $174 billion, up from $152 billion, growing at a 22 percent CAGR over the next few years. We are competing in all of the big AI deals and are winning significant deployments at scale.”
McGill said: ”We expect solid top line growth in the second-half of the year even as we continue to optimize our cost structure to enhance our competitiveness over the long-term.
“A big part of this optimization effort is leveraging AI to reimagine our business processes and drive higher productivity. To that end, in Q2, we took a $328 million charge for workforce reduction as we continue to position our business for the long-term.”
She added: “We’re using it [AI] to improve customer and team member experiences in sales, software development, services, content management and our supply chain. And in turn, we’re using our experiences to help our customers realize benefits of AI for themselves.”
It looks like fears about AI replacing human jobs are real.
Clarke said: “If you think about the use cases, the use cases are generally around large language models; small language models are coming. We’re already talking about technologies like agents and agent is a buzzword of our industry today, and we’ve just scratched the surface of how agents are going to help each individual’s productivity, help with specific skills, which provide an opportunity to take the millions of processes that are in businesses today and help automate them.”
Outlook
Although lagging currently, CSG has a positive outlook, according to Clarke, with: “modest commercial PC demand growth in the quarter …we continue to pursue profitable share focusing on commercial PCs, high-end consumer and gaming with our strong attach motion.
“We are optimistic about the coming PC refresh cycle, as the installed base continues to age, Windows 10 reaches end-of-life later next year and the significant advancements in AI-enabled architectures and applications continue.”
McGill said: “We expect [CSG] growth in the second-half of the year, particularly in the fourth quarter.”
The outlook of $24.5 billion +/- $500 million for the next quarter represents 10 percent Y/Y growth, with ISG revenues growing in the 30 percent region. The full fy2025 outlook has been revised upwards slightly, with revenues of $97 billion +/- $1.5 billion, again 10 percent higher than a year ago, with ISG up 30 percent or so and CSG flat. It looks highly likely that ISG revenues will overtake CSG revenues by the end of the fiscal year.
Nutanix beat the street with its latest results driven by a strong contract renewals performance and increased sales in its customer base, despite large deals taking longer to close.
Revenues in its fourth fiscal 2024 quarter, ended July 31, were $548 million, a 21 percent annual increase, and a loss of $126.1 million mainly due to a $106.4 million expense item. Full year revenues were $2.15 billion, up 15 percent from fy 2023’s $1.86 billion, and beating its guidance, but with a loss of $124.8 million. There was 22 percent year-over-year annual recurring revenue (ARR) growth in the quarter to $1.91 billion, strong free cash flow generation and a first full year of positive GAAP operating income ($8 million).
Rajiv Ramaswami.
President and CEO Rajiv Ramaswami stated: “Our fourth quarter was a solid finish to a fiscal year that showed good progress on our financial model with solid top line growth and sharp year-over-year improvement in profitability.” That is the non-GAAP profitability of $8 million, which Nutanix called an “important milestone.”
The revenue increase happened despite a problem, as Ramaswami explained: “Our land and expand business underperformed related to our internal expectations due to the longer-than-expected sales cycles.”
Financial summary:
Gross margin: 85.2% vs 83.7% last year
Free cash flow: $224.3 million vs year-ago’s $45.5 million
Operating cash flow: 244.7 million compared to $58.3 million a year ago
Cash, cash equivalents and short-term investments: $994 million compared to $1.7 billion at the end of the prior quarter
CFO Rukmini Sivaraman said: “The primary reason for the reduction in our cash balance was Bain Capital’s conversion of the 2026 notes, which we announced in June. We settled the conversion in Q4 by paying $817.6 million in cash and delivering approximately 16.9 million shares of common stock.”
Rajiv Ramaswami became Nutanix CEO in fy 2021, Since then Nutanix’ revenue growth rate has increased to a consistent high level – apart from a Q4 fy2022 blip.
The company gained 670 new customers in the quarter, its highest number for 3 years, attributed to new partners and sales programs. That means it now has 26,530 customers. It signed partnership deals with Dell, Cisco and Nutanix during fy2024. Ramaswami said: ”We see these partnerships as both expanding our addressable market and providing us with meaningful go-to-market leverage.”
William Blair analyst Jason Ader told subscribers that the current Nutanix-Cisco OEM partnership (Cisco Compute HyperConverged) contributed to significant new logo additions in the second half of fiscal 2024 with expectations for solid growth in revenue contribution in fiscal 2025.
Nutanix is finding that large deals are taking longer to close, Ader telling subscribers: “Nutanix’s largest win in the fourth quarter was a multimillion-dollar deal with a Fortune 100 financial services company that took 18 months to close, not unlike the eight-figure win announced last quarter that took over two years to close.”
Ader pointed out that: “While moving upmarket is a top go-to-market priority, Nutanix stressed that it remains very active in its traditional sweet spot at the lower end of the enterprise market. This segment of the market is particularly ripe for displacement of VMware as many of these smaller accounts have seen their VMware bills skyrocket under the new Broadcom pricing scheme.”
Nutanix noted that, where VMware is running on a 3-tier architecture, a switch to Nutanix and its hyper-converged infrastructure (HCI) necessitates a hardware refresh and this can push deals out to a scheduled hardware refresh date.
Ader also said: “Nutanix reached a Rule-of-40 score of 43 in fiscal 2024 (defined as revenue growth plus free cash flow margin), boosted by stellar free cash flow generation of $598 million.”
Ramaswami was asked about progress with AI and Nutanix’ GPT-in-a-box Gen AI offering in the earnings call. He replied: “I think a lot of initial interest in Gen AI has been in creation and training of LLMs, large language models, and a lot of that is being done in the public cloud and massive GPU farms. And we don’t play there fully.
“But on the other hand, we think the bulk of the enterprise opportunity in terms of how companies are actually going to use it, is potentially going to be on-prem because, at the end of the day, the Gen AI workloads, applications have to run wherever the customer data is. And in a lot of cases, sensitive customer data is either inside data centers or at the edges, and so a platform like our GPT-in-a-Box provides a very simple, easy-to-use, secure, way of running Gen AI applications.”
It is still early days for GPT-in-a-box adoption but the opportunity “for fine tuning, RAG (Retrieval, Augmented Generation) and for inferencing in terms of running these AI workloads close to the data in a private and secure way” is coming.
The outlook for the next quarter is $570 million +/- $5 million, meaning an 11.5 percent increase on the year-ago quarter. Nutanix is guiding full fy2024 revenues of $2.45 billion +/-$15 million in revenues; a 14 percent Y/Y increase.
DRAM and NAND fabricator SK hynix has revealed the industry’s first 16Gb DDR5 memory system using a 1c node – the sixth generation of the 10nm process, building on 1b.
The silicon slinger promised it will be ready for mass production of the 1c DDR5 “within the year” to start volume shipments next year.
The product improves cost competitiveness in manufacturing, compared with 1b, by adopting new materials and processes, said the South Korea-headquartered manufacturer.
SK Hynix 1c DDR5.
In May 2023, Samsung Electronics announced it had succeeded in mass-producing 1b-nanometer DRAMs for the first time in the industry. SK hynix then followed suit. Samsung said earlier this year it planned to start mass-producing a 1c DRAM by the end of this year – but no updates have been forthcoming, so SK hynix now seems to be ahead.
Double Data Rate 5 (DDR5) Synchronous Dynamic Random-Access Memory (SDRAM) promises to reduce power consumption and increase bandwidth, compared to its predecessor DDR4 SDRAM.
The operating speed of 1c DDR5 – expected to be adopted for high-performance datacenters – has been improved by 11 percent from the previous generation to 8Gbit/sec, said SK hynix. Power efficiency is also improved by more than 9 percent, we are told, as datacenters of course face increased power demands as a result of having to deal with AI workloads.
“We are committed to providing differentiated values to customers by applying the 1c technology, equipped with the best performance and cost competitiveness, to our major next-generation products, including HBM, LPDDR6, and GDDR7 systems,” said Kim Jonghwan, head of DRAM development at SK hynix. “We will continue to work towards maintaining our leadership in the DRAM space, to position SK hynix as a trusted AI memory solution provider.”
Earlier this month, SK hynix said it was developing a computational object storage system (OCS) with Los Alamos Nuclear Labs (LANL) to streamline analytics workloads. Object data is stored in Parquet files on NVMe SSDs, and when LANL wants to do a large scale simulation run, the data on the SSDs is preprocessed by the OCS to reduce the dataset sent to the analysis servers.
Veeam has widened its support for different platforms with the unveiling of its latest version of the Veeam Data Platform, including extra support for those migrating from VMware.
Since completing its acquisition of VMware at the end of last year, Broadcom has shifted its focus to larger customers. As a result, Veeam reports that its SMB customers are considering Proxmox VE as an alternative. Proxmox VE is an open source server virtual machine and container environment. Its toolset includes vzdump for backing up guest virtual machines and integrates with Proxmox Backup Server for centralized backup management.
Withv12.2 of the Veeam Data Platform, there is full backup support for Proxmox VE, protecting the hypervisor without requiring the management or use of backup agents. Users have flexible recovery options, including VM restores from and to VMware, Hyper-V, Nutanix AHV, oVirt KVM, AWS, Azure, and Google Cloud. They can also restore physical server backups directly to Proxmox VE for disaster recovery (DR) or virtualization/migration purposes.
“We see Veeam’s support for the Proxmox VE hypervisor as a valuable addition for organizations seeking open source flexibility and cost-effectiveness. It allows customers to stick with their preferred single platform for data backup and recovery while embracing the preferred hypervisor of their choice,” said Martin Maurer, CEO of Proxmox Server Solutions.
James Westendorf
“With the recent changes at VMware, we were forced to re-evaluate which hypervisor to choose for our datacenter – a situation nobody wants to be in. Veeam remains a critical partner for us as we look forward to being able to choose to move to Proxmox,” said James Westendorf, director of technical services at Lake Land College in Illinois, US.
“Using Veeam’s trusted tools for data migration will provide us with the peace of mind we need during these tumultuous times.”
With v12.2, there is also “enhanced” Nutanix AHV integration to protect critical data from replica nodes without affecting the production environment. Users can benefit from “deep” Nutanix Prism integration with policy-based backup jobs, “increased” backup security, and “flexibility” in network design, Veeam said.
The upgraded platform also provides backup for MongoDB, offering immutable backups, backup copies, and “advanced” storage capabilities, we are told.
In addition, there is expanded AWS support, allowing users to extend native resilience to Amazon FSx and Amazon Redshift through policy-based protection and “fast, automated recovery”. There is also wider Microsoft Azure support with native resilience to Azure Cosmos DB and Azure Data Lake Storage Gen2 for “resilient protection and automated recovery.”
“We are seeing customers evaluate a vast array of hypervisor and cloud infrastructure options to optimize both cost and functionalities, making data mobility more important than ever before,” said Krista Case, research director at analyst house The Futurum Group.
Earlier this month, analyst Gartner named Veeam the leader in market share for the global enterprise backup and recovery software market, displacing Veritas from the top position.
Pure Storage revenues grew in its second fiscal 2025 quarter, with hyperscalers and AI showing promise for the future.
Revenues in the quarter were $763.8 million, up 11 percent on a year ago and beating its $755 million guidance. Profit was $35.7 million, much better than $7 million loss last year. Pure said its discussions with hyperscalers to replace their core (disk) storage with Pure’s all-flash technology “continue to progress positively.”
Charles Giancarlo
Chairman and CEO Charles Giancarlo stated: “We were pleased with our Q2 revenue growth of 11 percent year-over-year. Pure continues to pick up market share, and outpace the industry, both in innovation and in growth.”
Quarterly financial summary:
Gross margin: 70.7 percent, flat year-on-year
Free cash flow: $166.6 million vs $46.5 million a year ago
Operating cash flow: $226.6 million vs $101.6 million a year ago
Total cash, cash equivalents, and marketable securities: $1.82 billion vs $1.2 billion a year ago
Remaining Performance Obligations: $2.34 billion, up 24 percent year-on-year
Pure’s revenue was made up of $361 million subscription services and $403 million product sales, with the year-ago numbers being $289 million and $400 million respectively. Subscription services are growing faster, at 25 percent, compared to product revenues, which had a 0.71 percent growth rate.
Pure noted strong FlashArray//E and FlashBlade//E capacity-optimized array sales performance. The FlashArray//C did well too.
There was 24 percent year-on-year growth in Pure’s Annual Recurring Revenue (ARR), driven by its EverGreen//One service offering. However, large EverGreen//One deals, those in excess of $5 million, took longer to close.
Pure gained some 261 new customers in the quarter, including two from the Fortune 500, taking its total customer count past 13,000.
Hyperscalers
It is continuing discussions with prospective hyperscaler customers. Giancarlo said: “Our lead prospect has advanced from extensive evaluation of our core technology to testing an integrated solution, and we have been engaged in detailed contractual negotiations for many months. We remain confident that we will secure our first hyperscaler design win by year end.”
CFO Kevan Krysler added: “Our most significant capital expenditures during the quarter were concentrated in engineering for new test equipment supporting key strategic growth initiatives, including our pursuit of hyperscaler infrastructure opportunities.”
CTO Rob Lee said: ”What we’ve done over the last many months is really move forward on our testing in phases from initial proof of concept to testing of that core IP to now extensive testing of really an integrated, think of it as a co-engineered solution. And as you’d imagine, this involves detailed performance, operational testing, so on and so forth.”
Giancarlo sees high potential with hyperscalers. ”The longer-term opportunity for Pure with hyperscalers is significant,” he said. “To provide a sense of scale, the top ten hyperscalers are projected to buy almost 70 percent of all disk drives, over 600 exabytes in total, this year alone. Because of Pure’s unique Direct to Flash technology, we can offer hyperscalers better performance, reliability, and power and space savings than hard disks, at a similar or better total cost of ownership.”
Krysler spoke in more detail about the hyperscaler opportunity for Pure’s Direct Flash Module (DFM) technology. ”They’re looking to make it an architectural shift in their infrastructure, whereby it would replace not only higher-performance workloads but lower-performance workloads, including disk, and once implemented, there’s no reason for it not to replace, for example, even their use of SSDs.”
Lee noted: “We think, in the long term, we’ve got opportunity to go provide value across the board with our Direct-to-Flash technology. But certainly, as an initial step, we’d be focusing on the replacement of the disk-based environments in that infrastructure.”
AI
Turning to AI, Giancarlo said: ”The AI market for data storage has progressed as we have consistently predicted,” with three identified opportunities. “First, storage for machine learning and training environments where Pure provides high-performance storage for public and private GPU farms.
“Our view is that in total, the total market for storage for large language type models – training models – is less than $1 billion a year currently.
”The second AI opportunity we foresee focuses on tailored storage for enterprise inference or RAG [retrieval-augmented generation] environments. Many, if not most, enterprises will use commercial LLMs or other models to operate on their own proprietary data in-house. These systems will use relatively small GPU environments to provide AI insight from their data. Pure is working closely with Nvidia on a number of vertical market offerings to satisfy this market.
“We continue to believe that our largest opportunity opened by AI is to address the siloed nature of enterprises’ existing data storage architectures. Current data stores sit behind application stacks and generally have neither the performance nor the connectivity to serve data directly for AI engines and analytics. Customers that are the most advanced in their AI investigations all acknowledge that data access and preparation are major barriers to AI deployment.
“Pure Fusion will allow customers to upgrade their enterprise storage to function as a storage cloud, simplifying data access and management, and eliminating data silos to enable easier access for AI.”
The Fusion offering unifies arrays and optimizes storage pools on the fly across structured and unstructured data, on-premises, and in the cloud. It also automates orchestration and workload placement.
CTO Lee said: “With our latest release of Fusion coming later this year, we’ll be able to deliver all of those capabilities that Fusion was designed for, being able to fully automate the management multiple environments, allow customers to manage their Pure Storage estate through policy declaration as opposed to individual operational steps and really step back and unify that as one pool of resources, one storage cloud, if you will. Later this year, customers will be able to take advantage of all these capabilities on all of their existing arrays and data storage estates.”
Giancarlo highlighted Pure’s electricity savings: “In a world where energy demands are soaring, the power savings of Pure Storage alone make the move from hard disks to Pure technology a smart choice for both hyperscaler and enterprise data centers. Businesses can grow their data storage and reduce their energy footprint with Pure on a platform that eliminates existing data silos and simplifies customers’ data centers with guaranteed service-level agreements.”
He added: “Energy and space savings generated by our Direct-to-Flash advantage are significant – we reduce space, power, and cooling requirements by a factor of 5 to 10 compared to hard disks. In a world of greater power demands and limited electrical supply, the savings on electricity alone provides a compelling incentive to switch from hard disks in both hyperscaler as well as enterprise data centers.”
William Blair analyst Jason Ader noted: “Investors remain on pins-and-needles with respect to a major hyperscaler win.”
If Pure does succeed in selling its flash storage to hyperscalers, perhaps it will start repeating its flash-will-replace-disk-drives message about which we have heard little recently.
Giancarlo is confident about Pure’s competitive strength. He told analysts: “We know we are gaining ground as our growing strength has forced competitors to intensify their efforts and mimic our messaging. It is clear now that legacy competitors in our market see Pure as the alpha competitor and have focused their messaging and strategies on us. We appreciate the attention and look forward to the competition.”
Krysler said: “We’re now the number two vendor of all-flash systems into the enterprise [and] firmly in that spot and only a few points behind the number one in that area … Competition is tough, but our lead in QLC remains.” Number one would be Dell.
The outlook for the third fiscal 2025 quarter is revenues of $815 million, up 6.8 percent year-on-year, with the full-year outlook unchanged at $3.1 billion, representing a 10.5 percent year-on-year growth rate. Giancarlo commented: “We believe that the storage market will be resilient in this IT economy, but we have yet to see a positive inflection.”
Wedbush analyst Matt Bryson commented: “Pure Storage ostensibly outperformed their prior outlook and guided Q3 above prior consensus revenue estimates, but the growth forecast wasn’t good enough in light of elevated expectations, particularly given other less-than-perfect metrics. We’d note that Pure Storage’s results contrast significantly with storage peer NetApp, which noted accelerating sales momentum and significantly stronger product growth for its July quarter.”
NetApp revenues are up, driven by all-flash array sales, making Q1 fiscal 2025 the company’s third successive growth quarter.
Revenues were $1.54 billion, which is 8 percent higher than a year ago, with a $248 million profit, a 66.4 percent increase from the previous year. NetApp generated $669 million in product revenues, a 13.4 percent increase, and its all-flash array annual run rate rose by 21 percent to $3.4 billion.
George Kurian
CEO George Kurian commented: “We started fiscal year 2025 on a high note, delivering strong revenue growth and setting records for first quarter operating margin and EPS … I am confident in our ability to capitalize on this momentum as we address new market opportunities, extend our leadership position in existing markets, and deliver increasing value for all our stakeholders.”
He added: “We are raising our FY25 outlook for both revenue and profit.”
Quarterly financial summary
Consolidated gross margin: 71.3 percent vs 69.6 percent a year ago
Operating cash flow: $341 million vs $453 million a year ago
Free cash flow: $300 million vs $418 million last year
Cash, cash equivalents, and investments: $3.02 billion
EPS: $1.17 vs $0.69 a year ago
Share repurchases and dividends: $507 million in stock repurchases
Kurian said NetApp experienced momentum in flash, block, cloud storage, and AI. He commented: “Both our capacity flash and block-optimized all-flash array families exhibited strong growth year-over-year, addressing an expanded TAM and driving share gains.”
He expanded on this in the earnings call, saying there were “really strong results across the board … in the block storage part of that portfolio, which is pure share gain against competition. We are demonstrating the price performance leadership against the high-end product of our competitors, as well as the price performance and feature set leadership against the midrange products of our competitors.”
In the performance flash block storage market, Kurian said: “We compete in that part of the market against frame arrays. It could be the Dell PowerMax or a large frame array from Hitachi [Vantara] or HPE.”
According to Kurian, QLC flash is being used in “the refresh or the migration of a very, very large 10K install base of hard drives, both ours as well as our competitors’, to that flash product portfolio … The total installed base of 10K drives is enormous. It was roughly, from a volume perspective, somewhere around 35 percent to 40 percent of the total storage market for a very long period of time. And so there’s a huge installed base to go refresh.”
Kurian noted strong performance in AI-led sales, saying: “While we believe the large opportunity for enterprise AI is still ahead of us, we are seeing good momentum today, with our AI business performing well ahead of our expectations. In Q1, we had over 50 AI and data lake modernization wins.”
He believes that NetApp is “in the early innings. It is when that inferencing trend, as well as large-scale generation of data come into play, that you’ll really see the inflection in data storage.”
“Inferencing is expected to be the preponderant majority of the storage market for AI and the enterprise AI landscape. It’s about 80 percent, maybe 90 percent of the total market. And RAG (retrieval-augmented generation) is expected to generate about 8x more data than the data that is fed into the RAG pipeline. So there’s a lot of new generation going to happen when and if these applications become mainstream.”
CFO Mike Berry commented about growth in general in the earnings call, saying: “Customers are prioritizing spend on strategic projects and so that part of our business continues to do really well … What we haven’t yet seen is large-scale datacenter refreshes, which would signal a broader base economic recovery and confidence in the business.”
He said customers want to use innovative AI tools in the public cloud, “where broader frameworks that combine databases, data warehouses, data lakes, together with AI models are progressing at a really rapid rate.”
But the necessary unstructured data sits in both on-premises storage as well as in the public cloud. “We are able to make that entire workflow much, much more secure and easy to manage, which is a part of the reason why we saw strength both in the data foundation for AI as well as in the cloud storage portfolio, where we’re seeing our ability to create a non-siloed, unified architecture come through for us.”
The public cloud portion of its revenues was $159 million, 3.2 percent higher than in the year-ago quarter. Kurian said: “Our highly differentiated first-party and hyperscaler marketplace storage services remain our focus and top priority. These services continue to grow rapidly, increasing roughly 40 percent year-over-year and performing ahead of our expectations at each of our hyperscaler partners.”
Berry said: “We expect cloud to return to a pattern of consistent growth through the rest of the year.”
Hybrid cloud revenues grew more strongly than public cloud, up 7.8 percent, to $1.38 billion. Flash arrays represented roughly 60 percent of NetApp’s hybrid cloud revenues. About 40 percent of NetApp’s installed base has adopted flash with the majority still using disk drives.
The Keystone storage-as-a-service offering experienced 60 percent year-on-year revenue growth.
Kurian thinks storage suppliers are doing better than broad systems vendors, saying: “The pure play storage players are outperforming the integrated system vendors. And so you look at a broad range of the integrated system vendors, they have struggled now for many quarters in their storage business, and it remains to be seen whether that is a strategic focus for them going forward.”
These are vendors, we understand, such as Hitachi Vantara, HPE, and IBM, but not Dell, with Kurian saying: “With GenAI, the two market players that have the installed base, Dell and us, are super well positioned.”
NetApp’s revenue outlook for the next quarter is $1.64 billion +/- $75 million and a 5 percent increase year-on-year at the midpoint; a lower increase than the current quarter’s 8 percent. Full year guidance is now $6.58 billion +/- $100 million, 4.9 percent up on fiscal 2024 revenues at the midpoint, and an increase from its previous $6.55 billion guidance last quarter.
EVP and CFO Mike Berry is retiring and a search is under way for his successor.
Startup QiStor claims it can accelerate storage engine key-value processing so much that two servers can do the work of 20, saving time, money, and reducing carbon emissions.
We were introduced to QiStor in February and have now learned more about its technology from CEO and co-founder Andy Tomlin, providing a clearer picture of its value proposition. It starts from the recognition that much database processing involves a software stack with a storage engine component.
OSS is Open Source Software
This piece of software talks across a network link, using a protocol like NVMe KV, to the storage drive hardware. But having applications talk through file systems to SSDs, which have Flash Translation Layers (FTLs) to map file storage IO requests to the drive’s block structure, is intrinsically inefficient. At heart, these layered constructs perform the allocation of space and the tracking of data locations on the drive. It’s simpler and more efficient to do this in one place through a key-value (KV) interface. Tomlin said: ”We’re doing it in hardware with full knowledge of how it’s actually working in the device.”
Andy Tomlin
QiStor is, in effect, replacing an SSD’s controller, or building an abstraction of a controller. Tomlin said: “I originally started the company to build a controller. But I abandoned that because just getting funding requires raising, ultimately, $100 million to build a controller, and no one was interested. So I figured out a cheap way to do this [and] we’re building an FPGA now.”
Tomlin says that key-value stores (KV stores) underlie NoSQL databases such as Aerospike, Amazon’s DynamoDB, Redis, Couchbase, LevelDB, and RocksDB. KV storage engine processing is computationally intensive and needs offloading from host server CPUs as it’s a waste of their resources and best handled by dedicated hardware.
Vector databases such as Milvus, Pinecone, and Vespa are used in large language model (LLM) processing for training and inferencing. Their operation can be accelerated by QiStor’s technology, reducing the necessary server infrastructure and speeding LLM responsiveness.
Tomlin says most large scale web applications today are built on key-value storage. He is firmly in the hardware acceleration and offload camp, strongly established by GPUs offloading graphics processing from x86 CPUs, and encompassing data processing units (DPUs) such as Intel’s IPU and Nvidia’s BlueField devices, smart NICs, and dedicated AI hardware. This kind of hardware acceleration is becoming mainstream.
He believes that hardware acceleration is needed for KV store software engines as well, and that relatively simple FPGAs can be used for this; Xilinx FPGAs, for example. Such FPGAs could be housed in database servers fitted in service provider datacenters, public clouds, or on-premises datacenters, and consumed through a Platform-as-a-Service (PaaS) facility.
Amazon EC2 F1 instances use FPGAs to enable delivery of custom hardware accelerations, and QiStor storage engine operations run on them as well as on Xilinx FPGAs.
QiStor is developing FPGA programs to provide functionality needed by the storage engine. The databases higher up the stack “talk” to the engine through existing access methods, Object-Relational-Mappers (ORMs) and libraries. It talks to the drive through a standards-based NVMe-KV interface. Tomlin says of the databases: “Most of the databases today already have plug-in storage engines … The idea is that the customer, the service, effectively, will run their databases, and they’ll just think they’re running on RocksDB as their back end.”
Tomlin says QiStor does not want to sell hardware. Other parties can do that, with QiStore producing the code that runs on or in it.
An initial task was to develop coding tools that can be used to write FPGA code, the hardware description language (HDL) describing the logic implemented on the FPGA. HDL code is converted by a synthesis tool into a netlist detailing the logic gates and interconnections needed on the FPGA, and this is converted into a bitstream, which is transferred to the FPGA. Now QiStor has developed its FPGA coding toolset, it can produce its code faster.
All this seems complex but it is more practicable and simpler than developing a storage engine-specific ASIC or other hardware, such as the Pliops XDP card.
Tomlin and QiStore are doing more than this. They are replacing an SSD’s block-based low-level structure with a KV storage system so that no traditional FTL is needed to map the KV data constructs to a target SSD’s block structure. He reckons this is inherently more efficient, with payoffs in eliminating FTL processing, simplifying garbage collection (recovery of deleted cells), and reducing memory requirements for SSD metadata. There is no need to use complex zoned infrastructure and the drives, with less write amplification, last longer.
Couple this with hardware acceleration and multiplicative performance advantages result.
A QiStor prototype demonstration ran with a 1 PB Aerospike database. When instantiated on x86 servers, this needed 20 x 60 TB servers, with 1,200 TB total capacity, delivering 753,000 write IOPS, and 3,017,000 read IOPS. The latency was 1 ms and the modeled three-year total cost of ownership (TCO) was $4.17 million.
This was set against just two servers with QiStor FPGAs and 1,000 TB of total capacity. They delivered 1.5 million write IOPS and 6 million read IOPS with a 250 μs latency. The modeled three-year TCO was less than $1 million.
Because fewer servers and SSDs were needed, the 58.2 kW of power needed for the pure x86 server setup was reduced to 5.8 kW for the pair of QiStor servers – a saving of 325 metric tons of CO2 per year.
It is becoming a truism that datacenters using GPU servers need lots of power, and the datacenter electricity budget left over for x86 servers, storage, networking, and cooling is much lower than in pre-GPU server days. This situation is likely to get worse. By lowering the amount of electricity needed to run petabyte-scale NoSQL databases ten times more efficiently, QiStor can make better use of that power budget.
Using QiStor technology can make NoSQL databases, including vector databases, run much faster on far fewer servers, saving cash, electricity, and carbon emissions. It’s the sort of technology customers, once they have adopted it, will see no going back. Many can’t wait to check out a beta test product.
Enterprise application lifecycle data manager and SAP data migration expert Syniti has been bought from its private equity owner by Paris-based Capgemini.
Capgemini, a global IT services business, with 340,000 staff in more than 50 countries, stated the acquisition “will augment the Group’s data-led solutions for clients across the globe, in particular large scale SAP transformations, such as the move to SAP S/4HANA.” Capgemini reported global revenues of €22.5 billion ($25.01 billion) in 2023 and is a vastly larger business than Syniti.
Aiman Ezzat
As we wrote in March “Syniti is a structured data lifecycle management supplier – with a focus on enterprise applications such as SAP and Salesforce – and not a file lifecycle management supplier.” It manages complex data quality, data migration and data governance initiatives for some of the world’s largest companies. The company’s Syniti Knowledge Platform is a cloud-based management platform and a SAP Endorsed App, sold as a SAP Solution Extension called SAP Advanced Data Migration & Management, by Syniti for 10 years or so.
The combination of generative AI and SAP expertise has made Syniti an attractive buy to Capgemini.
Aiman Ezzat, Capgemini CEO, said: “A strong data foundation is key to enable relevant, trusted and quality insights. It’s a top priority for clients as they look to unlock further value from their enterprise data thanks to generative AI.”
“Syniti and Capgemini share the philosophy that digital transformation will always require data transformation to drive critical business benefits.Syniti is a leader in its field and a trusted global partner for major SAP data migrations. Its global team will reinforce Capgemini’s data-driven digital core business transformation services.”
Kevin Campbell
Kevin Campbell, Syniti’s CEO, said: “Capgemini’s deep industry expertise and leadership in data & AI services highly complement Syniti’s data first approach in the market, to ensure that more organizations bring data to the forefront of their business and digital transformation projects.”
Bridge Growth Partners CEO Alok Singh said: “We are incredibly proud of the growth and development Syniti achieved. With the active support of our full team, including our highly experienced advisors, Syniti’s leadership team and employees transformed the company into a true market leader in data management that we believe will be an excellent strategic fit with Capgemini.”
Syniti, formerly known as BackOfficeAssociates, was started up in 1996, and raised $30 million from Goldman Sachs in a venture round in 2008. It was bought by private equity house Bridge Growth Partners in 2017 for an undisclosed amount.
Once bought by Bridge Growth Partners it stared expanding by acquisition to broaden its portfolio, buying HiT Software in 2010, and then Entota (2013), CompriseIT and Quadrate (2016), BackOffice Associates (2019), DataStewardshipPlatform (2019), Data Migration Resources, 360Science, Data Insiders and Sledgehammer all in 2021. Now it, with its more than 1,200 staff, has been bought again, and once more for an undisclosed amount.
The transaction is expected to close in the coming months, subject to regulatory approvals and other customary closing conditions.
