Michael Tricot and John Lafleur must be patting themselves on the back. In just under two years, their Airbyte startup has progressed from nothing to a $1.5 billion valuation and $181.2 million funding to develop its open-source Extract-Transform-and-Load (ETL) data analytics-feeding technology.
Airbyte started up in January 2020 with the aim of building open-source connectors from data sources to data lakes and warehouses. These would replace proprietary tools and also enable feeds from less popular data sources ignored by proprietary suppliers as being of too little value – long tail connectors. Progress was rapid; within 17 months, Airbyte claims it caught up with the ETL incumbents with 150 connectors running Docker containers and deployable in minutes on any platform.
Co-founder and CEO Michael Tricot said: “With the rise of the modern data warehouses, our mission is to power all the organisations’ data movement and doesn’t end at ELT. By the end of 2022, we will cover more types of data movement, including reverse-ETL and streaming ingestion.”
Michael Tricot (left) and John Lafleur (right)
Tricot, a former director of engineering and head of integrations at Liveramp and RideOS, founded Airbyte with John Lafleur. Lafleur is described as a serial entrepreneur of dev’ tools and B2B technology. A year after starting it up, and in a 12-month period, they took in $£6.2 million seed funding, then a $25 million A-round and, such was their progress, have just raised $150 million in a B-round. This B-round was led by Altimeter Capital and Coatue Management, also including Thrive Capital, Salesforce Ventures, Benchmark, Accel, SV Angel.
San Francisco-based Airbyte launched a compute time-based cloud service for its connectors in October. Its software enables businesses to create data pipelines from sources such as PostgreSQL, MySQL, Facebook Ads, Salesforce, Stripe, and connect to destinations that include Redshift, Snowflake, and BigQuery.
It also announced a community-based participative model in which it plans to share revenues with connector contributors. Airbyte expects to have a roster of 500 connectors by the end of 2022.
Jamin Ball, a partner at Altimeter Capital, provided a statement: “Airbyte has already made a huge impact in a very short period of time and has more than 1,000 companies lined up to take advantage of its Airbyte Cloud data service that is starting to roll out. There is tremendous market momentum on top of Airbyte’s disruptive model to involve its users in building the ecosystem around its data integration platform.”
Blocks and Files has never come across a startup until now, which, in less than 24 months, has gone from founding to a $1.5 billion valuation, and taken in $181.2 million across seed, A- and B-rounds in its second year.
In September, in the context of Fivetran raising $565 million in a single round, we talked about the notion of a funding frenzy for companies involved in sourcing and storing data for analytics. We calculate that 2021 saw a grand total of $6.3 billion in storage-related startup funding across 30 companies, with $5.0 billion of that going into data preparation and storage for analytics startups – quite the funding frenzy.
Here is yet more evidence, with Airbyte, that the investing community is seeing an astoundingly vast opportunity in this field. Soon, it appears, virtually every enterprise on Earth will be gathering data about its sales and operations for analysis.
VAST Data has doubled the storage density of its all-flash Universal Storage hardware to 1350TB of raw capacity, by supporting Intel 30TB QLC NVMe SSDs. That’s over a petabyte of effective capacity per rack unit at 5:1 data reduction.
This qualifies it to claim in a blog that, by providing twice the capacity in the same two rack units and the same power consumption, it cuts datacentre costs for space, power and cooling. It also has a Universal Power Control feature coming next year. This will enforce limits on system power consumption by intelligently scheduling CPUs to reduce peak power draw by 33 per cent as the power drawn by the system reaches a set limit.
VAST co-founder Jeff Denworth said: “With this announcement, we are eliminating all of the arguments for HDD-based infrastructure and making it even easier for customers to reach the all-silicon datacentre destination we first charted back in 2018.”
The blog explains: “The method behind Universal Power Control is pretty simple … hard drives draw more power when they’re moving their heads, SSDs draw much more power when accepting data writes than when processing reads. The write/erase cycle just uses more power than processing reads. With Universal Power Control, as the power drawn by the system reaches the power limit, the system starts reducing the number of active VAST Protocol Servers (aka CNodes), which reduces the rate at which the system writes data. Such an approach would be impossible with direct-attached shared-nothing storage architectures that tightly couple CPUs with storage devices.”
The blog compares compare a cluster of VAST 1350TB enclosures with Dell’s PowerScale A300 — an enterprise HDD-based archive system — and Pure’s all-flash scale-out FlashBlade systems. It says “A 4U FlashBlade chassis holds 15 blades, each with 52TB of flash delivering 535TB of usable capacity and consuming 1800W. … The PowerScale A300 nodes hold 15x 16TB hard drives per node, with four nodes in a 4U chassis that consumes 1070W. Assuming 20 per cent overhead for data protection, virtual hot spares, and the like, each chassis will deliver 768TB of useable space.”
It includes a graphic comparing VAST, Dell’s PowerScale A300 and the Pure FlashBlade systems:
The company claims that, by running at just 500 watts per petabyte (with the coming Universal Power Control), it is 11x more power efficient than Dell’s PowerScale A300 and up to 9x more power efficient than Pure’s Flashblade offering. It says it has a 5x datacentre density advantage over the PowerScale and FlashBlade products as well.
The blog also points out that HDD rebuild times are much longer than SSD rebuild times.
VAST says an unnamed large global financial customer has bought a more than 200PB VAST system to consolidate its performance and archive tiers from mostly HDDs into a single flash-based store.
The new 30TB drives can be mixed and matched in VAST systems with the existing 15TB SSDs in a single cluster.
Data protector Veeam has appointed a new CEO, Anand Eswaran, to take over from Bill Largent who continues as board chair. The company is showing all signs of contemplating an IPO.
Anand Eswaran.
Eswaran brings, Veeam said, extensive experience in developing new business models, executing on market expansion, and driving growth with an inclusive purpose-led and people-first culture. He was previously the president and COO of RingCentral, which supplies cloud-based communications and collaboration offerings for business. In its most recent results it reported ARR of $1.6 billion, a 39 per cent increase year-over-year — just what Veeam would like to do.
A Largent statement said: “To have someone with Anand’s experience on board will lead us into a new era of success, as we further accelerate into the cloud and evaluate the opportunity for Veeam to be a publicly traded company in the future.”
This is Eswaran’s first CEO gig, but he has a big company background. At RingCentral, he led Product, Engineering, Sales, Marketing, Services, Customer Care, Operations, IT, and Human Resources. Before that he looked after Microsoft’s Enterprise Commercial and Public Sector business globally, after having led Microsoft Services, Industry Solutions, Digital, Customer Care, and Customer Success — a global team of 24,000 professionals.
Prior to Microsoft he was a SAP EVP, head of its $5.4 billion Global Services business with 17,000 business process and technology professionals.
Veeam attained the number two position by revenue in IDC’s Data Replication and Protection Software Tracker for the first half of 2021, having overtaking Veritas and behind only Dell Technologies. It wants the top slot.
Eswaran commented: “Data is exploding and has become one of the most important assets for all organisations. As such, data management, security and protection are pivotal to the way organisations operate today, and failure to have a robust strategy can be catastrophic. Veeam has a unique opportunity to break away as we sit in the middle of the data ecosystem, with the most robust ransomware protection and ability to protect data wherever it may reside.”
A Veeam breakaway, meaning a jump in revenues, particularly ARR, would be a great prelude to an IPO. And an IPO would hopefully enable Insight Partners to get a decent payback for its $5 billion acquisition of Veeam in January 2020.
Cloud-based file collaborator Panzura has launched Managed Migrations — a service to move customers’ on-premises and hybrid NAS and object data to its cloud.
Managed Migrations includes dedicated engineering experts, start-to-finish implementation, and technical resources to move data, applications, and workloads to the cloud so customers can use its CloudFS global file system and Data Services — its SaaS data management offering. Panzura claims that a a lossless migration of data, applications, and workloads, and even mega-projects can be completed in weeks, or even days, instead of months.
James Seay, chief services officer at Panzura, offered a statement: “Panzura Managed Migrations expands our ability to help customers become future-ready as they tune up their IT infrastructure for the cloud revolution.”
The company says each customer gets a dedicated project manager, migration architect, and migration engineer. Migration starts with a roadmap including architecture, prototyping, and operational recommendations concerning the data sources, the destination, and network connectivity. Panzura then handles everything from hands-on execution, performance analysis, and real-time optimisation. Specialised automation capabilities help migrate data and workloads with speed and precision.
Panzura provides ongoing support for customers until their data and applications are fully migrated and workloads are in production. It provides staff training assistance so customers can run their own environment, or the Panzura Global Services team can run the environment on behalf of customers.
The company says Managed Migrations provides support for any hybrid multi-cloud migration through Panzura partnerships with Amazon Web Services (AWS), Microsoft Azure, Google Cloud Platform, IBM Cloud Object Store (iCOS), Wasabi, and Cloudian. The Managed Migrations service has high-speed file transfer capabilities, regional and cross-region offerings, with built-in data security and contingent compliance with HIPAA and other regulatory mandates.
Panzura claims its experts also provide recommendations to reduce a customer’s overall data footprint and incumbent cloud storage costs. The Panzura global file system consolidates unstructured data from all locations, after deduplicating and compressing, typically resulting in cost savings of up to 70 per cent.
This dedicated migrate-to-Panzura service will compete with generalised data migrators such as Datadobi and DIY tools like Windows Robocopy and the Linux/Unix Rsync facility. There is also Atempo’s Miria product which can migrate data and other dedicated destination migration services, such as Kioxia’s online data migration to its KumoScale flash box.
In effect Panzura is offering a white glove migration service to its own cloud and will even operate its cloud for customers. This may encourage Panzura competitors Egnyte and Nasuni to do likewise.
Japanese microcontroller embedded flash design company Floadia has developed a 7bits/cell — yes, an actual seven bits per cell — NAND technology that can retain data for ten years at 150°C, that will be used for a AI Compute-in-Memory (CiM) operations chip. Its use in SSDs looks unlikely.
The company’s announcement specifies analog data, but also says it is a 7bits/cell structure — which means digital data. Perhaps there is a “Lost in Translation” effect here. It also says that without its semiconductor design tweaks, a cell would only retain data for 100 seconds. How long its tweaked-design cell could retain data at room temperatures is not revealed.
This 7bits/cell technology is based on Silicon-Oxide-Nitride-Oxide-Silicon or SONOS-type flash memory chips developed by Floadia for integration into microcontrollers and other devices. Floadia said it optimised the structure of charge-trapping layers — ONO (oxide-nitride-oxide) film — to extend the data retention time when storing seven bits of data.
Floadia SONOS cell image
Its announcement says “the combination of two cells can store up to eight bits of neural network weights” which sounds odd. If one cell can store seven bits why shouldn’t two cells store 2 x 7 bits?
The CiM chip stores neural network weights in non-volatile memory and executes a large number of multiply-accumulate calculations in parallel by passing current through the memory array. Floadia says that makes it a good fit for edge computing environment AI accelerators because it can read a large amount of data from memory and consumes much less power than conventional AI accelerators that perform multiply-accumulate calculations using CPUs and GPUs.
It claims that its intended chip, despite a small area whose exact size is not revealed, can achieve a multiply-accumulate calculation performance of 300 TOPS/W, far exceeding that of existing AI accelerators.
SONOS is a charge-trap mechanism, trapping electrons in Silicon Nitride film. The retention life is controlled by optimising the thickness and film properties of oxide and/or Silicon Nitride films. The company’s website states: “SONOS is free from leakage of electric charge through defect or weak spots in the Bottom Oxide film caused by damage during Program and Erase operation, because trapped charges are tightly bonded with the trap site in Silicon Nitride Film.”
Standard MOS vs non-volatile MOS.
Also the “G2 cell consists of one SONOS transistor and two switching transistors placed adjacent to the SONOS transistor (see image above). This tri-gate transistor works as one Non-Volatile transistor operated by logic level voltage to switching transistors and high voltage only to the SONOS memory gate. Because of quite low programming current — pA order to each cell which is equivalent 1/1,000,000 of Floating Gate NVM — the wiring to the SONOS memory gate is treated like a signal line. And the power supply unit is able to be placed outside of the memory block — such as a corner area of the die. This unique feature of G2 provides LSI designers freelines of chip design and creation of new circuits combining logic circuits with Non-Volatile functionality.”
SONOS charge trap has much less leakage that a floating gate cell.
Floadia says the technology also uses Fowler Nordheim (FN) tunnelling technology to achieve extremely low power in program and erase operations, consuming 1/1,000,000 times current compared to conventional technologies using hot carrier injection for program/erase operation. It says: “G2 satisfactory supports operating temperature up to 125°C and 20 years of data retention life at 125°C.”
So we have 10 years at 150°C and 20 years at 125°C, which suggests retention periods at room temperature could be immense.
Whether this technology could be used in commercial SSDs is an interesting question and we’re asking a non-volatile memory expert, Jim Handy, about it.
Jim Handy’s view
Jim Handy of Objective Analysis told us “The first thing that I thought of when I saw your note was the company Information Storage Devices (ISD), which was acquired by Winbond in 1999. ISD made floating-gate chips that stored linear voltages to better than ±1 per cent accuracy (that would be around seven bits, if it were digitised). They got designed into lots of measurement equipment and into record-your-own greeting cards. My favorite application was an instrument to measure and record the stresses on bridges.
“Move ahead a couple of decades and you have Floadia doing the same thing, but a charge trap version targeting a different application: AI.
“It’s a good use of these technologies, and there’s a lot to be said in favour of making neural calculations in the linear, rather than digital, domain. AI can overlook noise and variations that add cost and complexity to digital implementations, and it’s trivial to perform multiplication and sums (linear algebra) in linear electronics at a high speed using very little energy.
“Floating gates and charge traps store voltages, and the use of MLC, TLC, and QLC has given developers a very good understanding of how best to manage that.
“As for using this in SSDs, that’s a different matter. If a flash chip didn’t need to run fast and be cheap then we may already have seen 7-bit MLC.
Fast: There’s a lot of sensitivity to noise when you’re trying to digitise multiple voltage levels, but if you average over time then you can manage that, if you have a lot of time. Who wants a slow SSD?
Cheap: You can store more precise voltages with a big charge trap than with a small one (likewise for floating gates). The bigger the bit, the fewer you get onto a wafer, so the higher the cost.
“But the whole point in going from SLC to MLC to TLC to QLC is to reduce costs. You wouldn’t do that by increasing the size of the bits.”
We had the opportunity to ask Cohesity VP of product marketing Chris Wiborg some questions about cloud data management, data storage growth, ransomware and Cohesity’s future. Enjoy reading his answers.
Cloud Data Management
Blocks & Files: Is it likely that the three main cloud suppliers move up stack and start offering their own data protection and then data management services on top of compute instances, storage classes and Serverless functions?
Chris Wiborg.
Chris Wiborg: One can never predict what areas CSPs may dive into next should they see the market opportunity as worthy of their attention. However, given the hybrid and multi-cloud nature of many customer deployments, most of the customers we talk to today prefer a third-party alternative.
Any given CSP being good at just protecting their own offerings likely will not be sufficient — similar to why even though database vendors have offered a certain level of data protection themselves for years … most customers have moved towards a provider whose support spans across many different sources (unstructured data, containers, VMs, other DBs, etc.) in the interest of consolidating silos as opposed to managing multiple flavors and instances of data protection and management.
Could one or more of the three big CSPs buy data protection and data management suppliers so as to acquire needed technology? For example, Google acquired Elastifile to gain file system technology.
It’s always a possibility. There certainly have been independent vendors that at times have shopped themselves around as an alternative exit strategy. Effectively integrating acquisitions not architected in a cloud era into a commercially viable cloud offering another discussion entirely.
Will external data protection and data management service suppliers — as opposed to in-house AWS, Azure and GCP supplied services — naturally gravitate towards multi-cloud vendor offerings?
Absolutely. The modern enterprise IT world is de-facto a hybrid, multi-cloud one — and likely to be so for the foreseeable future.
What other competitive differences might they have from in-house CSP-supplied services, like functional superiority? Would these be sustainable?
Well, one example would be a conflict of business interests. Let me explain: Would you expect Microsoft to be incentivised to play nicely with AWS (or vice versa) in providing robust support for workloads offered by the other? This is where more neutral third-party options likely will have room to navigate for some time to come: by providing the features customers need irrespective of the native hosting environment.
Data Storage/Management
We are facing a seeming relentless rise in unstructured data. Petabyte-level backup and archive data estates are becoming commonplace. Will these develop into exabyte-level estates and then zetabyte-level ones?
Given the continued exponential growth of data, yes. In fact, given this is the prediction season, this may be sooner than we imagine. I don’t want to put a date on it, but with 5G rollouts occurring globally now too, I would bet on it happening very soon.
Surely the cost of storing zetabyte-class unstructured data, some of it or even most of it with very low access rates, will lead to mass deletion exercises and a cap on data growth?
This is one reason that capacity efficiency with capabilities such as global (as opposed to, for instance, per-volume) dedupe is critical even today. And, yet another reason why data governance has an important role to play in data management going forward. What’s really worth protecting or keeping — and for how long? Businesses are still not on top of this.
This will increasingly play into the data management strategies of large organisations as they wrestle with the balance of the costs/regulatory requirement/SLAs equation as appropriate within their organisation.
Ransomware/Cybersecurity
Might we have a situation where we could view ransomware as an effectively solved problem?
Sadly, not very likely as cybercriminals continue to evolve their tactics and techniques. We’ve witnessed the shift from attacking production data only, to going after the backup data first, to now the rise in exfiltration and double extortion schemes — a progression we’ve seen unfold over just the past couple of years. And, sadly, the next extension of ransomware’s blast radius is likely being cooked up right now.
Are immutable and threat-cleaned backups along with fast recovery needed as normal business data protection practice for this to happen?
These should be table stakes for every organisation looking to harden their ransomware security posture with next generation solutions today.
How far off are we in your view of companies widely having this type of recovery as standard?
Within the past year or so, it’s the first thing on a customer’s lips when they meet with us and want to have us help them with. And what’s not surprising perhaps is that it’s not just IT Operations, but also SecOps now at the table helping set these requirements.
Cohesity focused
We have several large data protection suppliers, most of whom are engaged in developing SaaS offerings. For example, Cohesity, Commvault, Druva, Rubrik, Veeam and Veritas. They are also extending into cyber-security to fight ransomware. There appears to be a looming lack of greenfield customers.
Apart from the NewCorps out there, were there ever really any greenfield customers? Since we arrived with our initial offerings it has been about displacing more established legacy IT rivals. As our results show, we are adept at providing a next generation approach to age old IT problems. That is how we’ve been disrupting and gaining market share since the beginning, and we are very prepared for more competitive battles. When prospects see what we can do in a fair POC, our technology shines.
How will a company like Cohesity grow once it is competing for customers in a relatively static data protection and data management market against other large suppliers?
Vendors differ in what data management really means to them. Some are using those words while really just focusing on addressing the backup and recovery problem.
The legacy generation of data protection tools — which IT teams still grapple with to perform and manage routine backup and recovery functions — poorly scale to do more data management tasks such as file and object services, secure data isolation, governance and audit support, dev/test copy management, and running analytics to obtain insights.
More importantly, these tools are not just failing IT — they are also failing the business in terms of its online reputation, its operating efficiency, and its ability to use data as a strategic asset. If you look at tackling ransomware — most of the companies we are up against predate modern malware attack vectors.
Our view on this has always been more expansive, taking on other workloads. Our perspective is that the long game is really about supplying a platform and set of services that our customers need across the lifecycle of their data in a distributed (core/edge/cloud), yet not decentralised (single point of administration) fashion. Data protection is just one phase in that broader lifecycle of data management, but when I look around there’s a lot of Frankenstein’s monsters and not much built for the next generation of requirements.
Could Cohesity move into the very fast-growing analytics space?
We already are there at a “primitives” level — we supply building blocks for others to take a step further. We also view third-party extensibility as a key platform design tenet for us, and therefore believe that there is room to both grow and aggressively partner in this space — guided as always by where our customers lead us.
Data migrator and manager Datadobi announced the appointment of Charlie Collins as the company’s new channel sales director for the Americas. He’ll be responsible for developing and managing strategic plans with focus partners in North America and will report directly to Paul Repice, VP of Americas sales.
…
The Miami Dolphins NFL franchise is using Dell’s unstructured data storage and hyperconverged infrastructure systems to help expand the use of video for fan engagement, safety, and security for all events at Hard Rock Stadium. The Miami Dolphins standardised on Dell products for all applications, media asset management, safety and security, disaster recovery, data backup and virtualization. The Dolphins estimate the organisation has generated more than $1.2 million in cost savings, which was used to help fund a data recovery site.
Roger Cox (Sourced from Crunchbase.)
…
Roger Cox, a highly-regarded enterprise storage research VP at Gartner, passed away in the last 48 hours and will be much missed within the storage supplier community. He spent 22 years at Gartner, joining from Adaptec where he was a director for RAID Marketing.
…
The UK’s Barclays Bank has decided to go all-in with HPE GreenLake for its global private cloud. The GreenLake platform will host thousands of applications, more than 100,000 workloads, and support the bank in delivering an enhanced personalised banking experience for its customers. The workloads include virtual desktop infrastructure (VDI), SQL databases, Windows server and Linux. The migration from the legacy infrastructure to the private cloud is being performed by HPE Pointnext Services in partnership with the Barclays team. More details can be found on our parent website, The Register.
…
Hyve Solutions, which provides hyperscale digital infrastructures, has qualified Micron’s datacenter and enterprise-optimised 7400 E1.S SSDs with NVMe for its configurable Polaris 9219 platform. “We designed the E1.S form factor into our systems portfolio because it addresses several pain points for customers, such as flexibility, heat dissipation, optimal performance/power, energy savings through improved thermal cooling capabilities, and rack consolidation through storage per node improvements,” said Jay Shenoy, Hyve’s VP of technology.
…
ioSafe, which provides disaster-proof data storage devices, announced the new ioSafe 1520+ 5-Bay Network Attached Storage (NAS) device, a fireproof and waterproof device can protect up to 210TB of data. The ioSafe 1520+ suited for disaster-proofing data for the privacy-concerned, off-grid locations, small businesses, and departmental applications. It can also integrate with cloud applications and has optional disaster-proof expansion bays.
…
Kioxia America announced its CM6 and CD6 Series of PCIe 4.0 NVMe SSDs have earned VMware vSAN 7.0 certification, enabling them to be shared across connected hosts in a VMware vSphere cluster.
…
Lightbits Labs announced a partnership with Define Tech — a global independent software vendor for cloud-native computing in finance, life sciences, broadcast and media, and high-performance computing — to enable data-intensive workloads like AI/ML at scale. Lightbits supplies all-flash storage accessed across NVMe/TCP.
…
OWC announced the release of SoftRAID 3.0 for Windows. This includes support for RAID 5 on Windows 10 and 11. Users can create RAID 0/1/5 with Windows and RAID 0/1/4/5/1+0 (10) with Macs. Built-in OWC MacDrive technology lets you seamlessly move SoftRAID volumes between OSes. SoftRAID Monitor constantly watches your disks and alerts you if problems are detected. Volume validation ensures sectors can be read and parity is correct. Error prediction helps protect against unexpected failure.
…
File system supplier Qumulo’s Core product has been certified by Veritas Technologies as a qualified software platform for Veritas Enterprise Vault.
…
Samsung Electronics unveiled a lineup of automotive memory products designed for next-gen autonomous electric vehicles. It includes a 256GB PCIe 3 NVMe ball grid array (BGA) SSD, 2GB GDDR6 DRAM and 2GB DDR4 DRAM for high-performance infotainment systems, as well as 2GB GDDR6 DRAM and 128GB Universal Flash Storage (UFS) for autonomous driving systems.
…
TerraMaster D5 Thunderbolt 3.
China’s TerraMaster announced its D5 Thunderbolt 3 professional-grade RAID storage compatible with the latest Apple M1 chip-powered MacBook Pro, with M1 Pro and M1 Max and the latest macOS Monterey. It is compatible with the latest Thunderbolt 4 protocol and has five bays that can take 3.5-inch SATA disks and 2.5-inch SSDs — a total storage capacity of up to 90TB. It delivers 40Gbit/sec through its interface, and attains speeds of up to 1,035 MB/sec (test conditions: 5 HDDs, RAID 0 mode). The device is compatible with daisy chaining across multiple Thunderbolt 3 devices, and each one can be realized using the Thunderbolt interface.
Lee Caswell.
…
Lee Caswell, VP Marketing at VMware, has announced he’s off on a new adventure, after five years at VMware where he marketed vSAN and the HCI concept. Caswell joined VMware after a brief stint at NetApp and time at FusionIO and Pivot3 (the was a co-founder) before that. Ironically he was an EVP for marketing at VMware for a year before leaving to help startup up Pivot3. He said that, at VMware “We made it happen in HCI with 10x revenue growth to more than $1 billion, 30,000 new customers, and five consecutive years of MQ leadership. Now that was fun!” It looks like he may be joining a startup.
…
MSP360, a provider of backup and IT management products for MSPs and IT departments worldwide, has added Wasabi Object Lock immutable storage to the latest version of its MSP360 Managed Backup Service to help MSPs and internal IT teams protect cloud-based backups from ransomware, natural disasters, or accidental human error.
Nebulon originally developed a server add-in card called a Storage Processing Unit (SPU) which provided storage and other infrastructure management functions to a server, offloading its x86 processors. The SPU, which is managed over the public Nebulon ON cloud, became a Service Processing Unit and formed part of what Nebulon calls its smartInfrastructure. It has expanded to become a server setup and management tool with smartEdge and smartCore offerings.
David Vellante, chief analyst at Wikibon, said “The cloud is extending on-premises to both enterprise and edge datacentres. Nebulon’s two-server edge solution leverages public cloud infrastructure combined with the company’s innovative API-centric management and automation, making it an ideal solution for organisations that want a modern, cloud-simple experience for edge datacentres.”
Telcos with 5G towers, retailers, warehouses, medical clinics and cloud service providers with edge services are driving a trend for more than 50 per cent of enterprise-managed data to be created and processed outside the datacentre or cloud by 2025. This will include having high-availability servers at distributed edge sites with restricted physical space.
Such servers can be simple-to-deploy hyperconverged infrastructure (HCI) servers with three nodes. Three, so that one can function as a so-called quorum witness if there is a networking problem and the other servers become disconnected from each other while still being operational — the so-called split-brain problem. The quorum witness listens and responds to pings from the other servers, decides which one is correct and declares that it is the correct server for that edge site. Only that server can then write to shared storage.
Nebulon SPU components.
However, in space- and cost-restricted edge servers, having an actual physical third server present is not desirable. Craig Nunes, a Nebulon co-founder and COO, told us Nutanix “has the user put a quorum witness in a VM, and hosted on a different server from the two cluster nodes,” possibly outside the edge site. That needs a network switch added to the edge site system. He said VMware has a public cloud-based quorum witness and the user has to manage and maintain this. At scale, with hundreds or even thousands of edge servers to install, these alternatives become unwieldy.
Craig Nunes.
Nebulon has gone a few steps further and has its Nebulon ON facility provide quorum witness functionality across all of a customer’s edge sites, and have it set up automatically. The SPUs function as edge server setup infrastructure cards and don’t need a separate network switch — saving rack real estate, power and cost. By providing the quorum witness as a service in Nebulon ON, administrative tasks related to hosting and maintaining the quorum witness are eliminated. Server and software licence costs can also be eliminated.
Further, as Nunes explained, “the card offloads the data services work from the server CPUs. Generally, we’re going to be able to handle more VMs than an equivalent HCI config.”
Nebulon says its switchless, two-node edge system is an affordable high-availability alternative for enterprises looking to stretch a cluster across two racks for protection against partner node, power and top-of-rack (ToR) switch failures. Servers and the Nebulon SPUs are shipped to edge sites, installed, switched on, and then the SPUs set up the servers and the high-availability configuration with quorum witness functionality automatically. Its a ship, switch-on and manage-through-the-cloud system that scales readily to hundreds and thousands of servers.
Nebulon co-founder and CEO Siamak Nazari said “two-node cluster support is a requisite piece in Nebulon’s evolution toward developing the ultimate cloud-managed, on-premises infrastructure solution, ideal for both edge deployments and also datacentre projects, whether they be small or large.” Clearly Nebulon wants to move inwards from any edge site beach head.
FileCloud, a company started up in 2005 as CodeLathe which supplies highly secure file collaboration services (think Dropbox, Box, Egnyte, Nasuni and Panzura type services but more secure), has raised a $30 million Series A led by Savant Growth Fund with participation from Kennet Partners, plus a $10 million growth capital facility with Avidbank. Ray Downes becomes FileCloud CEO after a decade leading Kemp Technologies — a former portfolio company managed by Savant Growth’s investment team. Kemp co-founder Peter Melerud becomes FileCloud’s chief revenue officer. The backers want rapid worldwide market expansion. Eric Filipek, managing partner and co-founder of Savant Growth, will join FileCloud’s board of directors. Prior FileCloud CEO Madhan Kanagavel becomes president and CEO. So, after 16 years trundling along and gaining a customer base, Kanagavel and the board evidently decided new blood and funding is needed.
…
UK-based VDI specialist IMSCAD has launched a desktop disaster recovery offering aimed at providing customers with a backup not just of their data and infrastructure, but also their desktops. It gives all users a desktop to access from anywhere, in the event of disaster interrupting access to the existing on-premises environment.
…
NetApp announced that its ONTAP storage operating system is the first enterprise storage and data management platform to achieve Commercial Solutions for Classified (CSfC) validation for a data-at-rest (DAR) capability package. CSfC is a cybersecurity program led by the US National Security Agency (NSA) that validates commercial IT products that have met the highest level of encryption standards and security requirements for hardware and software systems. Recently, the NSA has recommended that federal agencies hosting secret or top-secret data utilise storage solutions that have been CSfC validated.
…
Pusan National University, Korea, researchers have developed a data restoration algorithm that uses correlations between existing information to restore missing data in an event log with a high degree of accuracy. The high accuracy of the algorithm guarantees its applications not only in current enterprises but also in future AI applications. The study has been published in IEEE Transactions on Services Computing. In event logs, events have attributes that are linked to other events in “single event” or “multiple event” relationships. In the former case, each attribute of an event corresponds to a unique attribute in another event. Based on this relationship, the researchers developed a Systematic Event Imputation (SEI) method that restores a missing value by simply referring to the available value it is linked to.
…
SK hynix 24Gbit DDR5 DRAM chip and modules.
SK hynix has shipped samples of 24 Gigabit DDR5 DRAM with the industry’s highest density for a single DRAM chip. Currently, DDR DRAM offerings mostly come in density of 8Gb or 16Gb. SK hynix’s chip was produced with 1nm technology that utilizes EUV process. It has improved production efficiency, increased speed by up to 33 per cent, and lowered power consumption by 25 per cent. The initial offerings on this product are set to be 48 Gigabyte (GB) and 96GB modules for supply to cloud datacentres. It is also expected to power high-performance servers for big data processing such as artificial intelligence (AI) and machine learning, as well as realizing metaverse applications, among others.
Scaleflux computational storage SSDs, with Supermicro’s FatTwin server, has been benchmarked to show it can double Redis-on-Flash SSD endurance — to three years or even more.
Redis is an in-memory NoSQL database and Redis on Flash (RoF) uses SSDs to extend the in-memory capacity by keeping the dataset’s “hot values” — keys and the data structure behind the keys, the Redis dictionary — in DRAM, and putting the “warm values” — the lesser-used portion of the dataset — in local Flash storage. When a value stored in Flash is accessed, it is promoted back into DRAM. In many cases, with a high Flash-to-DRAM ratio, Scaleflux says, the highly intensive workload can wear out SSDs after as little as 18 months.
Tong Zhang, co-founder and chief scientist of ScaleFlux, said “The small number of SSDs required by RoF deployments leads to some critical design considerations, as users require SSDs to individually offer high endurance mixed with excellent workload performance. Our collaboration with Supermicro addresses these needs and appeals to a variety of use cases. We look forward to astounding customers with the performance and cost efficiency.”
According to a benchmark white paper, Redis-on-Flash utilises RocksDB to manage the values stored in Flash. RocksDB employs a Log-Structured Merge-Tree (LSM-Tree) structure that makes insertions into the database very fast, but reads may require multiple accesses to storage to retrieve a record. RocksDB is suited for write-heavy workloads. A typical Redis-on-Flash system will write frequently to Flash but read only occasionally, and a practical recommendation is to maintain a 10:1 ratio of Flash to DRAM.
8-node Supermicro FatTwin server with 6x U.2 drive bays per node.
The ScaleFlux CSD 2000 product is an up to 8TB SSD fitted with QLC (4bits/cell) Micron flash and integrated compression/decompression compute engine hardware. Scaleflux says its performance is up to 6x faster than other NVMe QLC SSDs and its endurance is up to 4x better than other QLC SSDs. Customers can expect 3:1 or better compression ratios.
The compression hardware effectively amplifies the access bandwidth and also improves the random IOPS numbers. The CSD 2000 has either a half-width, half-height PCIe add-in-card or 2.5-inch U.2 design which matches Supermicro’s FatTwin half-width server design with configurable 4Ux 8-, 4-, and 2-node systems.
The FatTwin server is designed to enable two nodes per rack unit, doubling the compute density compared to full-width systems. Scaleflux says benchmarking data shows consistent, low-latency performance with RAM hit ratios as low as 60 per cent using two CSD 2000 devices per node. The CSD 2000’s compression capability enables SSDs to last three years or more in the write-intensive RoF environment.
The benchmark paper explains that “Flash latency is three orders of magnitude higher than RAM; nonetheless, running the workloads exclusively from Flash achieves 31 per cent of RAM performance for write (SET) and 56 per cent of RAM performance for read (GET). In both cases, the average latency remains below a 1ms threshold. The Flash storage layer provides an appreciable level of performance under worst case conditions while extending database capacity by nearly 10x.”
In fifteen days, WekaIO has hit its sales target for the entire quarter.
The high-performance, scale-out, parallel filesystem company put out this tweet yesterday:
It said it has closed three times as many deals this quarter as a year ago, and it’s still in the first month.
We are seeing a spectacular, extraordinarily fast-growing company in action here. Where does it lead? An IPO, or Weka getting acquired for a monstrously large number of dollars. Having stumped up just $67 million in funding, Weka’s backing VCs can be hoping for a positively majestic exit multiple.
COV-19 virus. CDC/ Alissa Eckert, MS; Dan Higgins, MAM - This media comes from the Centers for Disease Control and Prevention's Public Health Image Library (PHIL), with identification number #23312.
Cohesity has announced its Security Advisor application, which looks for malware holes in a customer’s Cohesity environment and adds an extra layer to its Threat Defense architecture.
Security Advisor scans the customer’s Cohesity environment, looking at security configurations, access control, audit logs and encryption framework. It provides a score that tells customers how they are performing against Cohesity’s best practice recommendations, and makes recommendations on how to address potential risks from bad actors — both internal and external — which can limit their exposure to cyber extortion.
Brian Spanswick, Cohesity’s chief information security officer (CISO), explained that “Enterprises use an array of tools to generate and manage data, and each tool has its own security settings — making it difficult to review every setting and control access across all their disparate technology. This lack of visibility and control leaves IT environments vulnerable to cyberattacks.”
Cohesity competitor Rubrik updated its data protection software with an in-house ransomware threat hunter earlier this month. Data protection suppliers are moving to add more active ransomware and other malware threat detection capabilities after an initial round of adding immutable backup vaults and clean backup scans.
Cohesity Security Advisor screenshot.
Security Advisor is part of the Helios SaaS management platform which provides in-flight and at-rest encryption, immutability, WORM, role-based access and multifactor authentication.
It complements Cohesity’s CyberScan application on the Cohesity Marketplace which can uncover cyber exposures and blind spots in a Cohesity production environment by running on-demand and automated scans of backup snapshots against known vulnerabilities.
If and when a Security Advisor scan shows a score that indicates an issue, the following actions can be taken:
Get more detailed information and recommendations;
Get a centralised view of all Cohesity cluster security settings across geographies, sites, and regions, through the Helios Dashboard;
Meet internal security assessments via a scorecard.
Security Advisor is available immediately to all customers globally at no additional cost. It can be accessed through the Helios Security Dashboard, under Security Advisor.
