Startup Observe has raised a $115 million B-round for its SaaS-based observability service following 171 percent ARR growth in its latest fiscal year.
Jeremy Burton
Observability and application performance monitoring (APM) software looks at applications – particularly cloud-native applications – and collects events and log data, activity metrics, component service invocations, and results. Observe adds meaningful business context to the raw basic data – such as customers and users, a shopping cart, a Kubernetes pod, a software build identifier, a problem ticket, and so forth. It brings this data to a single Snowflake repository and uses elastic cloud features for its analysis compute and capacity, helping issues get identified, understood, and fixed faster.
CEO Jeremy Burton explained: “Legacy monitoring and APM players, shackled by outdated architectures, are dead companies walking. As private equity or strategic acquirers strip them down for parts, Observe is taking a new approach designed for today’s modern distributed applications and massive data volumes. We’re thrilled to have investors who are thinking big and validating Observe’s approach in one of the fastest-growing segments in tech.”
A Gartner magic quadrant named the leading observability and APM vendors in 2023 as Dynatrace, Datadog, New Relic, Splunk, and Honeycomb. There were 14 other vendors mentioned in the MQ, but not startup Observe – it hadn’t registered with the Gartner analysts. Cisco has acquired Splunk while private equity buyers took Sumo Logic and also New Relic. The latter was bought by Francisco Partners and TPG at an equity valuation of about $6.5 billion.
Observe’s new investors include Sutter Hill Ventures, which led the round, and existing investors Capital One Ventures and Madrona plus a marquee new investor: Snowflake Ventures. They all put in the cash valuing Observe ten times higher – $325 million as we understand it – than the company’s $35 million Series A round four years ago. That included contributions from Michael Dell, Snowflake CEO Frank Slootman, and Pure CEO Scott Dietzen.
Observe funding history:
2017 – founded
2020 – $35 million A-round
2021 – $7 million venture round
2022 – $70 million A2 round
2024 – $115 million B-round
Total funding: $227 million
Recent growth stats are what pleased investors:
New average contract value (ACV) increased 176 percent year-over-year in FY24
Total Contract Value (TCV) increased 194 percent in FY24
Net Retention Revenue (NRR), which indicates stickiness of product, is 174 percent (industry-leading is considered 130 percent)
Mike Speiser, MD at Sutter Hill Ventures and a founder of Observe, elaborated: “Observe has … delivered a product that is architecturally different to everyone else. The incredible growth in ARR and NRR is testament to the fact that this new architecture is now paying off for their customers.”
Customers include Capital One, Reveal, and Top Golf. Mark Cauwels, managing VP, enterprise platforms technology at Capital One, provided a canned blurb: “Like many cloud-first organizations, our data volume continues to expand. Observe provides a centralized and pre-correlated data layer that meaningfully organizes telemetry data from many sources at scale, helping drive faster response times.”
Observe’s headcount increased more than 50 percent year-over-year and the biz is growing its sales organization, planning to expand its market presence in North America over the coming year. It expects to more than double the size of its business.
DataStax has integrated with Microsoft’s Semantic Kernel to help developers build retrieval-augmented generation (RAG) applications and vectorize data with Astra DB and Semantic Kernel’s extensible open source SDK for AI applications and agents. Any C#, Python, or full-stack application developers can build RAG apps and AI agents for their enterprise data using Semantic Kernel’s features for managing contextual conversations, multi-step functions, and connections with the Microsoft AI ecosystem. Semantic Kernel lets developers build agents that can call to existing code and be used with orchestration models like OpenAI, Azure OpenAI, GitHub Copilot, and Hugging Face. Key features of Semantic Kernel include semantic functions, chaining capabilities, planners, and connectors for various enterprise applications and data sources.
…
Data lakehouse supplier Dremio announced general availability of Dremio Cloud on Microsoft Azure, claiming lakehouse flexibility, scalability, and performance at a fraction of the cost of traditional data warehouses. Dremio Cloud on Azure dynamically and automatically optimizes resources and capacity based on usage. It has end-to-end TLS encryption and no customer data resides within the Dremio environment. Customers can control their data within their own Azure tenant, in storage services such as Azure Data Lake Storage (ADLS).
…
FileShadow is a SaaS service that connects user content from wherever those files are located – cloud storage, email, mobile devices, hard drives etc. – to a FileShadow Library. It can be used on a desktop, a browser, or a mobile device, and users can share their content with others through it. FileShadow has announced person detection and object identification in user images, allowing them to tag individuals (or other objects). It can help photographers organize images of a shoot, group images associated with an event, and label names of buildings or animals.
…
InterSystems has added vector search to its IRIS data platform to enhance its functionality for natural language processing (NLP), text, and image analysis tasks. This capability allows the IRIS software to manage and query content and related dense vector embeddings – particularly as it enables RAG integration to develop generative AI-based applications.
Michael Sotnick
…
Cloud file services supplier Nasuni has appointed Michael Sotnick as its SVP of business and corporate development. He joins Nasuni from Pure Storage, where he served as the VP of alliances and business development for eight years. Sotnick assumes responsibility for Nasuni’s global partnerships including Microsoft Azure, Amazon Web Services, and Google Cloud, and will also drive new strategic technology partnerships.
…
According to a source, a visiting NetApp senior exec told people at an Australian event last week that Dell could offload its storage business. Both Dell and NetApp declined to comment when asked.
…
Nvidia’s AI Enterprise software platform is generally available. It includes microservices such as NIM to speed up Gen AI application deployments, cuOpt for routing problems, and AI Workbench to create, test, and customize pre-trained Gen AI models and LLMs on a PC or workstation and infrastructure management enhancements. Check out the announcement blog for more information.
…
Media workflow NAS supplier OpenDrives has a forthcoming Atlas product and workflow-centric business model to be unveiled at NAB 2024 in Las Vegas, April 13–17. It will let Atlas customers adjust their capabilities based on their creative workflow requirements, with a straightforward process for upgrading or downgrading.
…
Samsung unveiled the expansion of its Compute Express Link (CXL) memory module portfolio and showcased its latest HBM3E technology at Memcon 2024 in Santa Clara, California. It introduced its CXL Memory Module – Box (CMM-B), a CXL DRAM memory pooling product that can hold 8x CMM-D E3.s devices and provide up to 2TB of capacity with up to 60GB/sec bandwidth and 596ns latency.
A CXL Memory Module – DRAM (CMM-D) technology, validated with Red Hat, has DRAM integrated with CXL to facilitate connectivity between an x86 CPU and memory expansion devices. Samsung also demonstrated its HBM3E 12Hi chip.
…
SIOS LifeKeeper Web Management Console screenshot
High availability supplier SIOS has a new web console for its LifeKeeper for Linux product. LifeKeeper protects apps such as SAP HANA, Oracle, and others from downtime and disasters. The LifeKeeper Web Management Console (LKWMC) features:
Simplified processes and intuitive interfaces to save time and reduce errors in configuration and ongoing management;
Set up Progress Tracking to monitor the installation process in real time;
Self-Help “Information Cues” simplifying configuration and management;
Self-Contained system simplifying the deployment process and reduces maintenance overhead, with no need for Java or X Window System installation on the server;
Language localization support initially available in both Japanese and English;
Simplified firewall management requiring only two TCP ports;
Access and functionality across a range of devices, including tablets and smartphones;
Support for leading platforms, operating systems, and enhanced SAP integration.
…
SMART Modular Technologies introduces its ultra-high reliability memory solution, Zefr ZDIMM memory modules, offered in both DDR4-3200 and DDR5-5600 form factors and available in mainstream densities. “The industry standard for defective parts per million, or DPPM, for DRAM modules is in the 3,000 to 5,000 range,” said Tom Quinn, SMART’s SVP of business development. “Typical Zefr ZDIMM module DPPM ranges from 200 to 300. This ultra-high reliability is valuable in critical data processing environments where uptime is paramount and downtime costs can quickly escalate.” Learn more at DDR4 ZDIMM and DDR5 ZDIMM microsites.
…
SQream, which supplies a scalable GPU data analytics platform, and Qantm AI, an advisory service for AI strategy, AI governance, and AI architectures, have announced a collaboration to provide SQream’s technology to Qantm AI’s customers.
…
StorPool Storage was named as one of Europe’s fastest-growing companies according to a Financial Times report, which ranked organizations by the highest compound annual growth rate in revenue between 2019 and 2022. The biz was 842 on the list, with a CAGR of 42.4 percent during the time period. StorPool was one of only 147 tech companies in the survey – and the only company from Bulgaria – to be ranked in each of the past three years.
Rick Scurfield
…
Synology says its Active Backup and C2 Backup offerings have achieved a milestone, safeguarding over 20 million SaaS user accounts, endpoints, servers, and VMs worldwide.
…
VAST Data has appointed its first CRO: Rick Scurfield, who comes after a short break from being NetApp’s chief commercial officer. Scurfield, who spent 20-plus years at NetApp, takes over sales from VAST president Mike Wing.
…
UK-based channel-first cloud and DR specialist VirtualDCS has launched CloudCover Guardian for Azure, powered by Veeam, which it claims is the world’s most comprehensive Azure backup service, protecting more than 250 configurable items in an established Microsoft 365 estate. It’s an alternative to Keepit, OwnBackup, and other Microsoft 365 SaaS backup services. CloudCover 365 offers more backup time slots than any other Microsoft 365 Backup solution in the market, where organizations can backup and protect data up to 12 times a day, providing a two-hour RPO window for business-critical data.
…
Western Digital has spread its 24TB HDD technology to its Red Pro NAS drives. The Red Pro has a 2–24TB capacity range, a 550TB/year workload, 600,000 load/unload cycles, 2.5 million hours MTBF, 6GB/sec SATA interface, 7,200rpm spin speed, 287MB/sec sustained transfer rate at the 24TB capacity level, and OptiNAND technology with iNAND UFS embedded flash drive for 20, 22, and 24TB capacities.
…
XConn Technologies announced the release of early production samples for its “Apollo” CXL 2.0 switch. It says the Apollo switch is the industry’s first and only hybrid CXL 2.0 and PCIe 5 interconnect product. On a single 256-lane SoC, the XConn switch offers the industry’s lowest port-to-port latency and lowest power consumption per port in a single chip at a low total cost of ownership. XConn’s Apollo will be the central component of a real-world use case demonstration of a CXL Memory Pool with Samsung during MemCon 2024, March 26–27, in Mountain View, California.
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.
As such, it occupies a different market sector from vendors such as Komprise, Datadobi, and Data Dynamics, whose focus is on unstructured file and object data and migrating it to lower-cost and slower storage as access rates diminish, not structured database record data. Yet the need for structured data lifecycle management is just as obvious as the need for file lifecycle management – possibly more so, as enterprise apps are often mission-critical with data placed in fast-access primary storage silos.
Syniti – like Komprise, Datadobi, and Data Dynamics – is a metadata scanning supplier, but its niche is enterprise app database metadata, with aspects related to business, technology, and application, rather than to file and object metadata.
Private equity-owned Syniti was founded in Boston in 1996 by Tom and Trish Kennedy as BackOffice Associates – an SAP-focused consultancy. BackOffice Associates developed information governance, data stewardship, and data migration software, such as its Syniti Knowledge Platform, for SAP, Oracle, and other ERP vendors.
Goldman Sachs invested in the business in 2011. In 2017, a majority stake in BackOffice Associates, with a $300 million valuation, was acquired from Goldman Sachs by private equity business Bridge Growth Partners. As a curiosity, ex-EMC CEO, chairman, and president Joe Tucci is a Bridge Growth Partners co-founder and chairman, and joined the BackOffice Associates board at the time. The BackOffice Associates name gave way to Syniti after this acquisition.
According to Bloor Research, Syniti now sells packaged cloud-native data management software “with about a third of its revenue coming from licenses rather than professional services.” Its more than 500 customers include American Airlines, Kraft, and Nestlé, and it has worked for more than 200 Forbes Global 2000 organizations. The business employs over 1,400 staff.
It has built more than 160 connectors to enterprise applications, with SAP a strong focus, meaning it can read and understand their metadata. This expertise is not something easily gained, and this is what partially prevents Komprise, Datadobi, and similar file-focused lifecycle management vendors moving into the structured data lifecycle world. This barrier works both ways: Syniti cannot easily move into the unstructured data lifecycle management market.
Chris Gorton
Syniti’s EMEA MD, Chris Gorton, told us in a briefing that one of Syniti’s first activities in a customer engagement is rightsizing. “For every application, [we ask] how have you managed that over time? The answer is typically ‘we haven’t.'”
This results in over-provisioned primary storage, with Gorton arguing: “There are SAP systems out there that are 60–70 terabytes. And when we do an initial assessment, we normally find that maybe six or seven terabytes of that is actually real operational data that they access.”
What is a typical trigger for customers to engage with Syniti? Gorton explained: “Because the technology vendors push them in that direction, because of the way they they acquire ERP software. It’s consumption-based. So they have to think, ‘well, hang on a minute, if I’m going to be charged on a consumption-based model, are we going to be storing and accessing data in the most efficient way?’ So, in a way, the ERP industry is forcing customers to be more cost-efficient.”
Syniti is not typically used as a one-shot exercise by its customers. It supports the view that data lifecycle management is an ongoing activity. Applications come and go, experience upgrades, move from the on-premises to the public cloud, and some get decommissioned. Data needs governance to make sure only the right people and apps can access it, and its retention meets compliance and privacy regulations. The data accumulated by decommissioned apps needs to be evaluated and, where appropriate, stored and made available to upstream apps.
This is another point of difference with file-based lifecycle management vendors: app decommissioning is not typically one of their focus areas.
Gorton explained that Syniti’s offerings are somewhat like IT plumbing. Syniti doesn’t, as we understand it, own the data. It provides data management and delivers data access to where it’s needed. For example, Gorton told us Syniti doesn’t see itself providing any GenAI-type copilots to analyze data. It sees its responsibility as delivering valid, cleaned, and compliant data to AI apps for them to analyze and process.
AI’s results, he suggested, will only be as good as the data it uses – and Syniti aims to provide the best quality data it can for Gen AI models to process.
XenData has added a Media Portal viewer to its on-prem and public cloud tape archive library so users can see previews of archived image and video files to select content for restoration.
The company builds the X-Series tape archive products, which have a front-end RAID disk cache. Previews, also called proxies, are low-resolution versions of the media file. They are created and written to disk when a media file is initially archived. XenData also supplies fully disk-based E-Series archive products. These scale out from one to four nodes, each with 280 TB of usable capacity. The Media Portal provides fast access to previews of archived files on tape that are otherwise slow to access and view.
XenData CEO Phil Storey said: “The Media Portal provides an excellent way for a user to know what content they have in their media archive. It is not a Media Asset Management system, but it provides a simple interface that provides the functionality that many of our users want.”
The Media Portal for Cloud runs on a physical or virtual Windows machine with a gateway to one or multiple public clouds, including AWS S3, Azure Blob Storage, and Wasabi object storage. It displays previews of all video and image files written via the gateway, including files stored on offline storage tiers such as AWS Glacier and the Azure Archive Tier.
The X-Series archive products can store offline cartridges externally as well as internally in the library racks. The Media Portal provides previews of all such external video and image content, together with the cartridge barcode information, which informs the user which LTO cartridges should be imported back into the library to be able to access the required content.
The Media Portal is sold as a software subscription and pricing starts at $980 a year. It will be available from May 2024.
Quantum stock price has been below the $1 mark on Nasdaq for more than 30 days and, as such, it was sent a letter by Nasdaq requesting it regain compliance.
The data manager has been dealing with an accounting issue regarding component pricing in product bundles that has affected its ability to file SEC reports for its second and third quarters of fiscal 2024. This issue has caused it to request a reporting date extension from Nasdaq to avoid delisting for not issuing financial reports. Now it has fallen foul of a second Nasdaq rule that threatens its listing status as well.
Quantum stock price history from Google Finance
The $1 stock price non-compliance letter arrived on March 19 and Quantum has asked for a hearing to present its plans to regain compliance plus a 180-day extension in which to do so.
CEO Jamie Lerner was appointed by Quantum in June 2018 following an October 2016 NYSE delisting threat for its stock price dipping below the $1 mark for 30 days. That delisting was avoided in April 2017 by a reverse stock split. Seven and a half years later, and following a June 2020 move to the Nasdaq stock market, it now faces a second delisting threat.
We may see a second reverse stock split to drive up the share price and regain Nasdaq listing compliance.
Separately, Quantum filed an SEC 8-K report detailing changes to its term loan credit and security agreement. This states that any financial proceeds from asset sales, aka “disposition of assets,” be prioritized to pay down existing debt. Quantum says this is consistent with its “broader efforts to prioritize certain financial and business projects targeting improvements to working capital, acceleration of new products and a more focused business.”
The recent i7 Raptor tape library product plan announcement is part of the “new products” aspect of this.
StoneFly has over 10,000 customers, but you’ve probably never heard of it. It supplies a unified SAN, NAS and object storage system has a three-layer virtual air gap system.
The biz supplies storage hardware and software presenting as an iSCSI SAN, Fibre Channel SAN, scale-out NAS and S3 object system, or all three with its StoneFly Fusion OS. This also powers hyper-converged infrastructure (HCI), backup and disaster recovery appliances which have the virtual air-gapping as a standard feature. Its Air-Gapped Vault runs in both on-premises deployments and also public cloud installations.
Mo Tahmasebi
CEO Mo Tahmasebi told us: “None of our customers have suffered from ransomware-corrupted data.”
StoneFly has an interesting and eventful past. It employs 125 people – 75 employees with contractors making up the rest. There are 25 salespeople, with Tahmasebi informing an IT Press Tour audience at its Hayward, CA, office: “I can close deals in 45 minutes.”
He said StoneFly has more than 10,000 customers, over 50,000 deployments and in excess of 2.5EB of storage capacity deployed. The customers include consumers, small-to-medium businesses, government agencies, universities, hospitals, financial institutions and Fortune 500 businesses. StoneFly products have been deployed in US Navy Littoral Combat Ships (LCS) and Virginia Class Nuclear Submarines.
Tahmasebi said: “We run our own datacenters, in Fremont, Oregon, and Massachusetts, using caged areas in co-los.” He also claims StoneFly has “high, very high” revenue and is “very profitable.” It spends virtually no money on marketing and gains customers by referral and word of mouth.
According to Tahmasebi, StoneFly kit is a third of the price of competitors like Dell and HPE, and lower cost than iXsystems. Its object backup target can be used with Veeam and costs less than an ObjectFirst Ootbi appliance.
Stonefly sells to Samsung and Toshiba in Korea and has undisclosed OEM partners.
Air-gapping
The air-gapping layers are at the repository, controller, and node levels.
Air-gapped repositories consist of one virtual storage controller connected to two target storage repositories. One target repository is network-facing, always accessible and available to user-groups, applications, etc. The second target repository is air-gapped, detached, and isolated.
Air-gapped repository diagram
StoneFly says air-gapped repositories can be deployed on popular hypervisors and public clouds. Policies can be defined to automatically turn-on (attach/connect) and turn-off (detach/disconnect) air-gapped repositories.
Air-gapped controller diagram.
The node-level air-gapping involves the chassis power being turned off so the node becomes invisible from the network point of view. The virtual air-gap then becomes a physical air-gap.
Stonefly history
Present day StoneFly came into being through an acquisition after the original company – founded as StoneFly Networks an iSCI storage supplier in San Diego in 2000, and owning the www.iscsi.com website – ran into not-making-enough-sales trouble in 2003. It was VC-funded with Allen Yuhas as its CEO from 2002 onwards by which time it had raised $12 million in funding. StoneFly competed with LeftHand Networks (bought by HP in 2008), EqualLogic (bought by Dell) and Intransa (crashed in 2013) in the IP SAN market.
Dynamic Network Factory (DNF) bought StoneFly Networks’ assets in 2006, paying $205,000 for them, changing the name to StoneFly and operating it initially as a San Diego-based DNF subsidiary. By then StoneFly had taken in a total of $34 million in funding developing its IP SAN products and had sold around 175 systems over a 13-month period.
That year, 2006, SAN interconnect hardware supplier QLogic became an OEM for Stonefly, as did Permabit. StoneFly integrated Permabit’s Albireo deduplication technology into the StoneFusion Network Storage Platform software, and Permabit was bought by Red Hat in 2017. QLogic was bought by Cavium in 2016, and Cavium was bought by Marvell in 2018.
DNF was founded by President and CEO Mo (Mohammed) Tahmasebi and CFO Macy Tafreshian in 1989. It started as a US subsidiary of Japanese conglomerate CSK Electronics, and focussed on storage software and hardware from 1998 onwards. DNF was spun off into a privately held entity in 1990, with Tahmasebi and Tafreshian self-funding it.
DNF has thousands of customers and major partners include Microsoft, Intel, Western Digital, Seagate, and Red Hat. The customer list includes include UC Berkeley, MIT, John Hopkins, the Department of Defense, the Federal Aviation Administration, Lockheed Martin, Bank of America, Citibank, Wells Fargo, Fujitsu, Nordstrom’s, Toshiba, PG&E, Safeway Corporation, Looksmart, CNET, and LSI.
Prior to starting DNF, Tahmasebi:
Worked at Sun Microsystems as an EnterpriseEngineering manager from 1982 to 1990;
Co-founded Pars which ran from 1990-1998 before being acquired;
Started up Tzone Corp & Research in 1997 and it was acquired in 2000;
Founded Computer Training Academy in 1997 which was acquired in 2000;
Set up Salesforce1 in 1997 – it was acquired in 2014;
Founded Trundl in 2014 for it to be acquired in 2019.
StoneFly CFO Macy Tafreshian worked with Tahmasebi as a VP at Pars, and was a co-founder at Salesforce1.com and DNF.
Snowflake finds GenAI analysis of data in its cloud data warehouses is rising and wants to encourage it.
The company has published a Snowflake Data Trends 2024 Report in which it looks at AI-related change in more than 9,000 users’ activities over the past 6-12 months.
Jennifer Belissent, principal data strategist at Snowflake, said: ”Conversational apps are on the rise, because that’s the way humans are programmed to interact. And now it is even easier to interact conversationally with an application. We expect to see this trend continue as it becomes easier to build and deploy conversational LLM applications, particularly knowing that the underlying data remains well governed and protected.”
The report says: “Organizations are picking their models, creating more complex LLM applications, making AI more available to a wider range of users, and reaping the benefits of a unified data platform. There has been a lot of hype around the transformational potential of AI, but judging from what we’re seeing in the Data Cloud, the frenzied fanfare is beginning to materialize into concrete results.”
Four trends are identified in the report, the first being greater use of machine learning (ML) functions, which grew 67 percent between July 2023 and January 2024. This “indicates the enthusiasm for, and the utility of, these ‘democratizing’ functions.”
But Snowflake admits via the report: “These are early days, and of course that growth surge starts from a relatively small initial point, but we’re excited to see sustained and growing interest in tools that put more and more of the power of advanced AI into the hands of less-technical users.”
LLM-powered apps being developed by Streamlit developers are on the rise, the report says. Streamlit is an open source Python language library that can be used in Snowflake to process data in Snowflake’s data warehouse. “Within the Streamlit developer community, between April 27, 2023, and Jan. 31, 2024, we saw 20,076 unique developers work on 33,143 LLM-powered apps (this includes apps that are still in development).”
Not all the developers are corporate ones, but most are: “In a survey of 1,479 respondents, nearly 65 percent said their LLM projects were for work.”
Python language use in the Snowflake Snowpark (non-SQL code development) environment is rocketing. It ”grew considerably faster than both Java and Scala in the last fiscal year: Python grew by 571 percent, while Scala grew by 387 percent and Java grew 131 percent.” Snowflake says Python is an AI-friendly language.
The third trend is the rise of GenAI chatbots compared to single-text-input LLMs, which Snowflake says don’t allow “refinement through iterative text input.” The report notes that Streamlit chatbot development is catching up to single-text-input LLMs, although the latter still dominates in terms of weekly use:
Snowflake’s fourth trend is increased use of its native apps. These are deployed inside Snowflake’s data cloud and the native app framework was previewed at the end of June last year. Comparing July 2023 to January 2024, Snowflake says it saw “311 percent growth in the number of Snowflake Native Apps published … 147 percent growth in installation/adoption of these applications … and usage of these apps grew 96 percent.”
The report adds that “given the choice, users want to build applications within their data platform – where the data is – rather than export copies of the data to external technologies.”
Users also want to manage access to their data in Snowflake with data governance functions restricting access to authorized users, the report states. “The use of all data governance functions has increased by 70 to 100 percent. As a result, the number of queries of protected objects has increased by 142 percent.”
Get a copy of the report here (registration required).
After running its IPO, Astera Labs saw its shares, initially priced at $36/share, soar by 72 percent on March 20, closing at $62.03. This surge in share price valued the company at $10.4 billion, taking into account stock options and restricted share units. Astera will hire another 200 engineers and look at M&As as a way to grow its engineering team.
…
The Backblaze product portfolio has been added to Carahsoft’s National Association of State Procurement Officials (NASPO) ValuePoint contract. This enables Carahsoft and Backblaze to provide the company’s cloud storage to participating states, local governments, and educational institutions. This certification is augmented by the recent addition of Backblaze to Carahsoft’s NYOGS and OMNIA-approved vendor lists.
…
Data manager Datadobi launched StorageMAP 6.7. A blog says it has extended REST API extensions:
Add or configure file or object servers, streamlining the setup process for large data management projects
Dynamically adjust server throttling, allowing for precise control over performance and resource utilization
Retrieve real-time status updates of ongoing data management jobs, including critical information such as status and error counts
v6.7 also integrates replication functionality into StorageMAP. It now encompasses both N2N (NAS-to-NAS) and O2O (object-to-object) replication capabilities, consolidating all replication functionalities in a single facility.
…
Data recovery services supplier Fields Data Recovery has launched a Partnership Program offering:
Access to state-of-the-art data recovery technologies and tools
Comprehensive training and support to enhance partners’ proficiency in data recovery processes
Flexible partnership models tailored to meet the unique needs and objectives of each partner
Marketing and sales assistance to help partners promote their data recovery services effectively
Priority access to Fields Data Recovery’s expert team for technical assistance and consultation
China-based SmartX has released its SMTX Kubernetes Service 1.2 (SKS), a production-ready container management and service product. Updates include enhanced support for high-performance computing scenarios such as AI, multiple CPU architectures, and optimized management and use of container images. With these updates, SKS 1.2 helps users apply AI scenarios in containers based on easily created Kubernetes clusters.
…
Swissbit says its new PS-66(u) DP SD and microSD cards have proven robustness and industrial suitability, and are certified CmReady by Wibu-Systems. This enables the integration of CodeMeter technology for software protection and license management. The use of the Swissbit cards as additional license containers within the CodeMeter ecosystem offers a way for developers of embedded and IoT devices to protect their software and monetize their licenses.
…
Taiwan-based TeamGroup has launched the MagSafe-compatible PD20M Mag Portable SSD, and the ULTRA CR-I MicroSD Memory Card Reader. The PD20M Mag External SSD is weighs 40g and measures 70 x 62 x 8.2 mm. It stores up to 2 TB, has a USB 3.2 Gen 2 x 2 Type-C interface, is compatible with the iPhone 15 Pro, and supports Thunderbolt interface ports with a maximum transfer speed of 20 Gbps. The ULTRA CR-I MicroSD Memory Card Reader supports UHS-I interface slots and USB Type C. The maximum data transfer speed is 180 MBps when used with TEAMGROUP MicroSD memory cards. It is meant for use with smartphones or other mobile devices.
…
Veritas announced enhancements to Backup Exec, which has more than 45,000 small and midsize business (SMBs) users worldwide. The updates include:
Malware detection – powered by Microsoft Defender, can be used to scan both VMware and Hyper-V backup sets at any time or prior to recovery
Role-based security – limits access to data based on a user’s specific role. In the event an account is compromised, hackers can only corrupt the small volume of data associated with that specific user’s account
Faster backup and recovery – optimizes protection performance with forever incremental backup of VMware and Hyper-V. Protection of virtual machines is now faster with parallel backup of multiple virtual disks and the included ability to recover virtual machines instantly
Backup Exec is available as a subscription service and can be installed in ten minutes or less, we’re told. Once installed, completing the first backup takes only five minutes. Find a Veritas-certified partner or sign up for the Veritas Backup Exec free 60-day trial.
…
UK disaster recovery specialist virtualDCS has launched CloudCover Guardian for Azure, which protects more than 250 configurable items in an established Microsoft 365 estate. Configurations covered by CloudCover Guardian for Azure span user accounts, access privileges, and unique security groups across popular applications such as SharePoint, Teams, OneDrive, Exchange, and Entra ID (formerly Azure Active Directory). Microsoft does not currently offer a backup service for these configurations. The system captures every vital configuration, meaning a Microsoft 365 environment can be proactively managed from a single console for a complete, clean recovery after an incident.
The company has also launched a “Clean Room” service for customers that need to restore their systems in a sterile and isolated environment in the event of a ransomware attack.
…
GenAI represents a Fourth Industrial Revolution, says Wedbush financial analyst Daniel Ives. He told subscribers: “With a March Madness-like atmosphere at the Nvidia GTC Conference with the Godfather of AI Jensen leading the way, its crystal clear to us and the Street that the use cases for enterprise generative AI are exploding across the landscape. The partnerships and spending wave around AI represents a generational tech spending wave with now Micron the latest example of beneficiaries from the 2nd/3rd/4th derivatives of AI just like we saw from Oracle a week ago. The ripple impact that starts with the golden Nvidia chips is now a tidal wave of spending hitting the rest of the tech world for the coming years. We estimate a $1 trillion-plus of AI spending will take place over the next decade as the enterprise and consumer use cases proliferate globally in this Fourth Industrial Revolution.”
The GenAI hype is getting relentless. Is it a dotcom-type boom and bust, though? Ives asserts this is “NOT a 1999 Bubble Moment.”
Real-time in-memory database vendor Redis is acquiring Speedb and its data storage engine.
A storage engine is used to write data to and read data from storage drives and is written in low-level code. Speedb, a 2020 startup, has developed an API-compatible plug-in replacement for the RocksDB storage engine to accelerate RocksDB-using applications. Redis is a distributed NoSQL database, keeping data on storage drives for persistence, and promoting it to DRAM when needed. It can be used as a database, including vector functions, cache, RAG, streaming engine, or message broker. Redis Enterprise Cloud, a real-time managed database-as-a-service, is offered on AWS.
Rowan Trollope
Redis CEO Rowan Trollope said: “Modern app development and most recently generative AI are fundamentally changing how enterprises value and use their data, and as a result, the demands on developers and data platforms are intensifying. Whether it’s regarding development cycles or user experiences, a lot of this pressure revolves around speed.
“Acquiring Speedb takes Redis beyond RAM, and opens the door for us to serve an integral role in powering today’s most exciting and innovative applications, and supporting the developers that build them.”
The Speedb storage engine fits right into the Redis software stack, and can be used to accelerate its storage drive IO. In fact, Speedb was integrated as the default storage engine in the enterprise auto-tiering functionality launched in Redis 7.2. The GenAI angle is pertinent as Redis has developed its low-latency vector database for AI workloads such as semantic search, LLM semantic caching, and session management. Redis has developed a Redis Vector Library (RedisVL), a streamlined client that enables the use of Redis in AI-driven tasks.
Adi Gelvan in the foreground
Speedb co-founder and CEO Adi Gelvan said: “As we join the Redis team, we’re excited to scale our technology to thousands of organizations across the globe, and play a critical role in powering the emerging applications that will shape the future.”
A Redis statement says: “Taking full advantage of the advancements in SSD storage and transfer rates that Speedb leverages, Redis will serve the full spectrum of performance and cost requirements for enterprise applications where DRAM is not required – empowering developers across the Redis ecosystem to utilize Redis in more use cases.”
The acquisition cost has not been revealed but is unlikely to exceed seven figures. Speedb has taken in a relatively low $4 million in funding across a November 2020 seed round and a non-equity assistance raise in November 2023.
In the immediate future, Redis is developing Redis Community Edition 7.4, with the addition of hash field expiration and client consolidation features. It will formally release the first version of its Speedb-integrated product later this year with Redis 8.
Scalar tape library supplying Quantum has made an early announcement that an i7 Raptor library is coming, designed for hyperscale customers’ AI workloads.
Update: Quantum clarifies i7 Raptor status versus i6H. 25 March 2024.
Its present Scalar line has three products: 25–400 tape slot i3 with 1–24 half-height drives, 50–800 slot i6 with 1–24 full-height drives, and high-end i6000 with 100–14,100 slots and 1–192 full-height drives. There is also a scale-out i6H product, dating from 2022, which is for web-scale companies – hyperscalers in other words – and based on a Redundant Array of Independent Libraries (RAIL) design providing RAID-like protection and performance. The drive numbers and tape slot counts for the i6H have not been revealed, but we think its constituent 48RU-high racks each have 24 drives and 800 tape slots with in-rack robotics. There are many similarities between the i6H and i7.
A canned quote from Bruno Hald, VP of secondary storage at Quantum, states: “Large enterprises need a low cost, highly secure archival storage system that acts as the backbone of private and hybrid clouds, and creates data lakes to fuel AI models and initiatives. As the pioneers in tape technology, we used our expertise and input from our major hyperscale customers to develop the most advanced tape solution built for these emerging use cases.”
Quantum claims the i7 RAPTOR – it capitalizes the name – delivers the highest storage density of any tape library on the market, and offers unique anti-ransomware features like Tape Blocking, which prevents a library robot from moving tape cartridges from their slots via software commands.
The i6H also supported this logical tape block feature.
The i7 features AI-driven and automated predictive analytics that can monitor drive-media interactions, predict robotic failures, and gather system data to learn and help analyze and improve performance.
The i6H also featured predictive analytics, but now it’s 2024 so it’s mandatory to use the term “AI”.
Quantum has an ActiveScale object storage system and the i7 integrates with this, offering an S3 interface, enabling integration into S3-compatible AI work streams.
The i6H also integrated with the ActiveSale repository.
Quantum claims that, because of the i7’s high storage density, fewer libraries are needed to achieve maximum storage capacity, saving on datacenter floorspace. We’re also told that the product has the industry’s lowest power consumption and employs sustainable materials and processes throughout its lifecycle, from manufacture to delivery to operation to maintenance to disposition. It’s a good green story.
There are so many i6H and i7 similarities that we think the i7 may be a productized i6H, and have asked Quantum if this is the case – also asking for drive type and number data and the tape slot minimum to maximum range. A Quantum spokesperson told us: “The Scalar i7 is a completely new library with a new design from the i6H, and it’s also much denser.”
Quantum picture of i6H library.
We asked how many tapes it supports and the answer was: “Quantum will release more details on slot numbers when the product is generally available but anticipate it will be the densest library on the market.”
How many bulk load slots (I/E) does it support? “It will support bulk import/export of tape cartridges, and the numbers will be available closer to GA.”
What is its Capacity on Demand (CoD)? “It will have CoD, and Quantum expects it to be in 100 slot increments.”
What number and type of tape drives does it support? “LTO-9 FH drives at launch. LTO-10 will be added when it becomes available.
Please share an image of the product. “We will have product images to share later this year.”
The i7 Raptor is available as either an opex model – with monthly, usage-based payment schedules – or an as-a-service model, which provides more flexibility and scalability. It will be generally available early in 2025, with units going into customer testing and certification by late this year.
Comment
Quantum has announced an intention to ship a tape library product in ten to twelve months time, but tis is not an actualproduct announcement. The i7 Raptor has an as yet undefined number of tape slots, tape drives and bulk load cartridge numbers and there is no product image available.
Chad Sakac and fellow execs at Regatta are getting ready to discuss the OLTP and OLAP (Online Transaction Processing and Online Analytical Processing) database for developers that, we undertsand, is scheduled to be released later this year.
Also known as Captain Canada after a famous costumed zipline descent into an astounded 2017 Dell EMC World conference hall, Sakac previously served as Dell’s VP for VMware Tanzu in and among the many senior roles he held at the company. He left in March 2021 to take a sabbatical and by October that year he had moved to an unnamed start-up which was still locked in stealth mode.
That fledgling business was Regatta where Sakac became Vice President for Go To Market. The company was founded in 2019 by Boaz Palgi, CTO Erez Webman, and VP R&D Eran Borovik. The three have been involved in Topio (acquired by NetApp), Xtremio (acquired by EMC), Storwize (acquired by IBM), and ScaleIO (acquired by EMC).
There is some information available about its funding, with Pitchbook mentioning a 2020 a $12.5 million A-round and a 2022 $45.8 million B-round. There are some 50 engineers developing the product. OLTP is generally used for real-time processing of online transactions, with OLAP used for multi-dimensional, multi-dataset analysis. OLAP databases may get their data from OLTP databases via Extract, Transform and Load (ETL) procedures which can be complex and take time.
Chad “Captain Canada” Sakac soars down to the stage at the 2017 Dell EMC World Event in Las Vegas
Regatta is ACID (Atomicity, Consistency, Isolation, Durability) compliant, meaning database operations are completed properly or do not complete at all, ensuring the database is in a consistent state. In other words, a bank’s OLTP database will ensure that money has either been paid out from your account or has not. There is no in-between state.
Regatta is pitched as being linearly scalable, running across a cluster of nodes either on-premises or in the public cloud, and analytic queries execute against up-to-date transactional data. OLAP queries do not interfere with OLTP operations. The database can, Regatta says, be dropped in as a Postgres replacement, having Object-Relational Mapping (ORM), Client Libraries, and ANSI SQL support. It is a high-availability product, we’re told, with auto-healing and geographical multi-site disaster recovery.
Chad Sakac
Sakac blogged about Regatta’s database in October last year, saying developers will like it because it won’t become complex and messy like some existing databases, urging them to try it out.
In a separate blog post, he wrote: “Regatta can store multiple types of data as well as perform transactional, analytical, and high-ingress workloads – and it can do them simultaneously. That eliminates the operational headaches of ETL, as well as the complexities of managing silos of data.”
Regatta is running a webcast on March 28 to introduce itself with Palgi, Webman, and Sakac speaking. We understand that Regatta will announce availability of its product later this year.
Flash memory and storage producer Micron topped its own financial guidance for Q2 of fiscal ’25 ended February 29, on the back of ever growing demand for IT infrastructure that trains and powers artificial intelligence.
Revenues of $5.8 billion were 57.6 percent higher than a year ago and Micron brought home $793 million in net profit versus a $2.3 billion net loss in the corresponding quarter of the prior financial year.
Sanjay Mehrotra, president and CEO, said in statement: “Micron delivered fiscal Q2 revenue, gross margin and EPS (earnings per share) well above the high end of guidance. Micron has returned to profitability and delivered positive operating margin a quarter ahead of expectation.”
Micron is climbing out of its revenue trough faster than it expected – thanks to the Gen AI boom.
Wedbush financial analyst Matt Bryson commented: “We believe Micron earnings commentary ticked every box in the investor checklist (and then some).”
Mehrotra thinks Micron is one of the biggest beneficiaries in the semiconductor industry of the multi-year opportunity enabled by AI. This is due to memory and NAND demand for AI training and also inference both in the datacenter and at the edge.
In Micron’s results presentation, server demand is said to be driving rapid growth in HBM, DDR5, and datacenter SSDs, which is tightening leading-edge supply availability for DRAM and NAND. This is resulting in a positive ripple effect on pricing across all memory and storage end markets.
The price rises are unlikely to slow down or stop during Micron’s financial year, it indicated. Mehrotra said in prepared remarks: “Our current profitability levels are still well below our long-term targets, and significantly improved profitability is required to support the R&D and capex investments needed for long-term innovation and supply growth.“
Financial summary
Gross margin: 18.5 percent vs -0.7 percent last year
Operating cash flow: $1.22 billion vs year-ago $343 million
Free cash flow: -$29 million
Cash, marketable investments & restricted cash: $9.72 billion vs $12.1 billion last year
Diluted EPS: $0.71 vs -$1.12 last year
Dividend: $0.115/share
DRAM pulled in $42 billion, up 53.8 percent annually, while NAND revenues of $1.6 billion grew even faster at 80.7 percent. Mehrotra said: “We see strong customer pull and expect a robust volume ramp for our mono-die 128 GB product, with several hundred million dollars of revenue in the second half of fiscal 2024. We also started sampling our 256 GB MCRDIMM module, which further enhances performance and increases DRAM content per server.”
He discussed HBM3E memory, used in Nvidia GPUs, saying: “Industry-wide, HBM3E consumes approximately three times the wafer supply as [DDR5] to produce a given number of bits in the same technology node; a 3:1 trade ratio. With increased performance and packaging complexity across the industry, Micron expects the trade ratio for HBM4 to be even higher than the trade ratio for HBM3E.”
That will limit the fab capacity available for DRAM and, we reckon, have an upward effect on prices. Micron is planning to increase its manufacturing capacity overall. Announced projects in China, India, and Japan are proceeding as planned and potential US expansion schemes assume CHIPS Act grants for Idaho and New York fabs in its in its capex plans for FY24.
Micron commenced volume production and recognized its first revenue from HBM3E in fiscal Q2 and has now begun high-volume shipments of its HBM3E product. Micron expects 12-high HBM3E will start ramping in high volume production and increase in mix throughout 2025.
End markets
Compute and Networking: $2.1285 billion, up 59 percent year-over-year
Mobile: $1.598 billion, up 69 percent
Embedded: $1.111 billion, up 28 percent
Storage: $905 million, up 79 percent
DRAM and NAND demand has sent the compute and networking BU’s revenues up sharply, with the mobile and storage BUs benefitting as well. The embedded sector, less affected by AI, is growing more slowly.
The PC market is starting to grow slowly, in the low single digits, after two years of double-digit declines. Mehrotra said momentum was developing for “next-generation AI PCs, which feature high-performance Neural Processing Unit chipsets and 40 to 80 percent more DRAM content versus today’s average PCs. We expect next-gen AI PC units to grow and become a meaningful portion of total PC units in calendar 2025.”
Micron said smartphone unit volumes in calendar 2024 remain on track to grow at a low to mid single digit rate. It expects so-called AI phones to carry 50 to 100 percent more DRAM compared to non-AI flagship phones today. Mehrotra said: “The Honor Magic 6 Pro features the Magic LM, a seven-billion parameter large language model, which can intelligently understand a user’s intent based on language, image, eye movement and gestures, and proactively offer services to enhance and simplify the user experience.”
B&F expects the other DRAM and NAND fabbers and SSD suppliers to report similar revenue uplifts in the coming weeks.
Outlook
Micron expects DRAM and NAND pricing levels to increase further throughout calendar 2024 and is guiding for record revenue and much improved profitability in fiscal 2025. It says we are in the very early innings of a multi-year growth phase driven by AI as this technology ay well transform every aspect of the business world. On that basis, its outlook for the next quarter is for revenues of $6.6 billion, give or take $200 million, a 76 percent increase on the year-ago Q3.
