Home Blog Page 233

Sprawling, fragmented data is a ticking time bomb. Here’s how to defuse it

By
Joseph Martins
-
Illustration of ransomware ghost coming out of someone's locked PC to demand money

Paid Feature The data landscape is often described in pyrotechnic terms. Data growth is explosive, even meteoric, while ever smarter analytics and AI means its potential is incendiary.

This vivid imagery is, perhaps, justified, given that almost 40 percent of organizations expect their data to grow by half or more every year, according to IDC research for HPE. But data is also fragmented, with the same research showing that just 10 per cent of organizations have the entirety of their digital infrastructure residing on-premises.

At the same time, the threat to data has increased enormously. The prospect of cyber attackers – or disgruntled insiders – exfiltrating data, whether PII Data or crucial IP is bad enough. But the threat of having data locked up by ransomware means companies are simultaneously prevented from carrying on with their business is existential.

Ransomware has certainly caught the imagination of tech leaders, with 27 per cent of respondents citing it as their most prevalent threat in the same IDC research.

Research conducted by the security vendor Cybereason reveals 81 percent of organizations are highly or very concerned about the risk of ransomware attacks. And so they should be. If current trend lines are anything to go by, ransomware is expected to attack at business, consumer or device every two seconds by 2031, up from 11 seconds in 2021. And if the attack is successful? The average downtime is 21 days.

With all these pressures, it perhaps shouldn’t be a surprise that half of organizations had suffered an unrecoverable data event in the previous three years.

So, should we be worried about the sort of adjectives used to describe organizations’ existing data protection? Words such as fragmented, aging, complex. And, perhaps worst of all, secondary.

As HPE’s data protection product marketing manager Ashwin Shetty explains, in this data rich, rapidly cloudifying world, organizations must often rely on tooling that is typically designed for a decade earlier. Moreover, “earlier backup and recovery was considered secondary storage and complex to deploy.”

This has meant that both operations – and technologies – really are fragmented. “You have multiple point solutions for backup, disaster recovery as well as archives. No single vendor has been able to unify this and offer it as a single solution.”

Just maintaining a multitude of point solutions on-prem sucks up time in management and maintenance. This becomes even more complex when customers inevitably begin to extend their operations to hybrid cloud architectures, because much of the existing data protection tooling was designed for an on-prem world. Things are further complicated if a team’s desire to run mixed workloads means it requires additional media servers and backup targets.

This is a stark contrast to the situation in cloud, which offers customers a more unified, agile experience, which is often impossible to replicate on-prem. However, this comes at a price and in any event, many applications and many data pools simply can’t be moved off-prem.

Lastly, there is the people factor. Traditionally, the team responsible for data protection and recovery could have been very distinct from the security team. Now, given the prevalence of ransomware and the almost inevitability that organizations will be hit by it, it is important that the two disciplines are, if not united, at least working in concert.

Seamless tooling

So how should data specialists begin to rethink retooling data protection for the modern era? There are three important questions that organizations need to ask themselves, Shetty explains: “How can I secure my data against ransomware attacks; how can I recover from without disrupting any of my business operations; and finally, how do I simplify backup and recovery operations?”

“Customers need to have a solution that can be as seamless as configuring your required RTOs and RPOs, and then you have the right blend of disaster recovery, or backup and recovery operations set up for you. Implementing a data protection solution has to be as simple as a few clicks.”

So, simplicity and ease of implementation are among the key elements of HPE’s GreenLake for Data Protection. HPE GreenLake brings a consumption-based, cloud experience to the management, provisioning and optimization of infrastructure, and services, whether it is on-prem, in a colocation facility or, indeed, in the cloud. The offering is underpinned by HPE’s InfoSight AI-powered optimization engine.

HPE’s Greenlake for Data Protection offers two key new services; HPE Backup and Recovery Service, which aims to modernize data protection and bring operational simplicity across hybrid cloud; and disaster recovery, ransomware recovery and continuous data protection, courtesy of Zerto, a Hewlett Packard Enterprise Company which HPE acquired last year.

Shetty adds that customers aren’t locked in with HPE GreenLake. “It allows customers flexibility to choose the backup ISVs of their choice. It can work with ISVs like Veeam, Commvault, Cohesity and others.”

But, he continues, “what it helps you to do is make management and maintenance of your infrastructure either on-prem or in the cloud much easier, much simpler, because the operation is managed not by the customer but by HPE GreenLake.”

According to Shetty, storage efficiency is a key differentiator of HPE Backup and Recovery Service, thanks to the firm’s StoreOnce Catalyst dedupe and compression technology. “Our internal testing shows that we are up to five times more efficient than any other backup service protecting VMware workloads.”

Policy-based orchestration and automation lets you set up the protection of your virtual machines (VMs) in a few simple steps, in less than five minutes. All backups will then run automatically so that you can be sure you can recover your data when you need it.

Managed through a unified cloud native console, this modern, flexible approach to backup automatically retains local snapshots for instant data restores, performs local on-premises backups for rapid data recovery, and utilizes cloud backups for cost-effective, long term data retention.

Using HPE StoreOnce Catalyst makes this service efficient at protecting backup data from cyberattacks like ransomware, because it keeps data hidden from attackers. Ransomware can’t infect and encrypt what is inaccessible to it. HPE Backup and Recovery Service takes advantage of this feature, creating backup stores which are not directly accessible by the operating system. Backup images are made inaccessible to ransomware, thereby ensuring data backup security and enabling reliable data restores.

Likewise, the initial focus is on VMware, which continues to account for the vast majority of corporate workloads. “But we will start supporting other workloads quite soon.”

It’s all in the mix

When it comes to mixed workloads, Shetty adds, “one of the key things with the backup and recovery service is you don’t need additional gateways or additional media servers to start protecting your VM workloads. Everything is done as a part of this service.”

As for setting SLAs and RPO/RTO targets, the equation is simple. Companies want minimum downtime and minimum data loss. “That’s where Zerto comes into play.”

With just a few clicks, you can recover your multi-VM applications to seconds prior to an attack, according to Shetty. Applications and all their VMs are recovered as one single, crash-consistent unit. Using built-in orchestration and automation, you can easily failover an entire site in a few clicks – without disruption, he says.

Using Continuous Data Protection (CDP) with its always-on replication and journaling technology, you can recover entire sites and applications at scale, with seconds of data loss. Simply select a checkpoint that’s seconds before an attack, and resume operations.

One of the headline claims of DPaaS is that it will allow customers to tackle ransomware “head on.” That might seem almost provocative given the nature of the threat and the people behind it. But Shetty explains the aim is to make ransomware attacks a “non-event.”

“You have near zero data loss in the process, and you have near zero downtime, so you don’t have to spend time recovering the data. You don’t have to spend time restoring your systems back after an attack. And the data is tested to make sure it’s a clean restore.” Customers can easily perform failover and backup testing quickly, without disruption. Using an on-demand sandbox, you can test when you need to and receive automated reports.

None of this denies that cyberattacks in general and ransomware in particular are very real threats and it is inevitable that companies will come under attack. But the ultimate impact, says Shetty, “depends on how quickly you’re able to restore your data and resume your business operations”.

Making a ransomware attack a non-event might be another way in which data protection fails to grab the sort of superlatives data growth does. But, let’s face it, that’s probably how we’d all prefer things played out.

Sponsored by HPE.

Veeam stops sales in Russia over invasion of Ukraine

By
Chris Mellor
-

Backup supplier Veeam has declared it will no longer sell its products in Russia in response to the country’s military invasion of Ukraine.

Veeam CEO Anand Eswaran issued a blog post about “Veeam’s Actions in Support of Ukraine” in which he confirmed: “Veeam has paused all sales in Russia.”

Anand Eswaran

He wrote: “As a company that prides itself on a culture of equality, teamwork, innovation, and action, we cannot stand by while those values are taken away from the Ukrainian people. The need for freedom, stability and peace outweighs the need to sell software.”

Veeam is making access to all its data protection products “along with full production level support free to all business and government entities in Ukraine, for the duration of this conflict as well as the recovery phase.”

It has set up an employee contribution match to registered charitable organisations up to $1,000,000 and is donating $250,000 to organisations focused on providing medical aid, helping with children and families’ basic needs, and assisting with the escalating refugee crisis.

Veeam is also supporting the use of employee volunteer days to help make a difference at this time.

Some readers may consider it’s time now for all storage sector suppliers to stop sales to and support in Russia.

Background

The company was founded in 2006 by Andrei Baranov and Ratmir Timashev and set up its HQ in Switzerland, with software research and development in a St Petersburg office, the Russian city on the Baltic coast. 

Andrei Baranov (left) and Ratmir Timashev (right)

The founders focused the company on backing up and restoring VMware virtual machines faster and better than anyone else and saw sales soar for more than two decades. By 2020, annual revenues had surpassed $1bn and its worldwide customer count had passed 365,000.

Veeam was bought by Insight Venture Partners for $5bn in 2020. American William Largent, then EVP for operations, was promoted to CEO and board chair. Veeam became a US-led and headquartered company. Baranov and Timashev stepped down from the board. Timashev styles himself on LinkedIn as a contracted consultant at Veeam.

Anand Eswaran was appointed CEO in December last year with Largent remaining as the chairman.

Nevertheless, Veeam has a Russian aspect to its corporate identity and the CEO has moved quickly to distance the company from Russia’s invasion of Ukraine.

Kaspersky

Another prominent Russian company, security supplier Kaspersky, has adopted a neutral stance with CEO Eugene Kaspersky tweeting: “We welcome the start of negotiations to resolve the current situation in Ukraine and hope that they will lead to a cessation of hostilities and a compromise. We believe that peaceful dialogue is the only possible instrument for resolving conflicts. War isn’t good for anyone.”

That stance is not getting a lot of positive comments.

Faster SQL engine and new metastore for Dremio’s ‘data lakehouse’

By
Chris Mellor
-

Data analyser startup Dremio, which wants to kill off data warehouses, has launched a data lakehouse cloud, a Sonar SQL engine, and an Arctic lakehouse metadata store using Apache open source projects.

A data lakehouse is a portmanteau term combining “data lake” and “data warehouse” and meaning, in Dremio’s sense, running data warehouse SQL functions on a raw data lake. There is no need for extract, transform and load (ETL) processes to pre-process the data and load it into a data warehouse such as ones from Teradata and Snowflake.

Dremio CEO Billy Bosworth issued a statement: “In the past, companies had to weigh the pros and cons of data lakes vs data warehouses. We’ve eliminated this tradeoff by providing a frictionless and infinitely scalable service that delivers the combined benefits of both.

“Dremio Cloud is the world’s first lakehouse platform that was built from the ground up for SQL workloads, including mission-critical BI (Business Intelligence).”

Dremio Cloud is a free data lakehouse with data persisted in Apache Iceberg format tables using Apache Parquet’s columnar data format. Iceberg is a high-performance format for petabyte-scale analytic tables and is said to bring SQL table simplicity to big data. Parquet includes compression and encoding, and can handle complex bulk data. 

An enterprise edition of Dremio Cloud offers greater security with custom roles, enterprise identity providers (e.g. Okta), and enterprise support options. 

Sonar and Arctic

Dremio says Sonar, a SQL engine powered by Iceberg, is more than twice as fast as its prior SQL engine. The idea is that customers can save money by using 50 per cent of the cloud infrastructure they would otherwise have needed. Sonar can work on data in S3, in the Dremio Cloud or in other data lakes, and supports SQL Data Manipulation Language (DML) insert, update, and delete operations directly on the lakehouse.

Dremio claims that, as data is automatically optimised in Amazon S3 as it is being changed, customers no longer need to fall back to a data warehouse when data is rapidly mutating.

Tomer Shiran, Dremio founder and chief product officer, said “Organizations can easily analyse data that resides outside the lakehouse (e.g. in relational databases) and perform live and accelerated joins across multiple sources.”

The Sonar software uses metadata and management services such as Glue, Amazon’s serverless data integration function, or Dremio’s Arctic service to connect to and work with source data. Dremio says Sonar obviates the need for ETL processes and data lakes; work directly on data lakes instead with the familiar SQL language.

Arctic is in public preview with support for a variety of lakehouse engines, including Spark, Flink, Presto, Trino, and Dremio Sonar.

Dremio is offering a forever-free edition of Dremio Sonar and Dremio Arctic on Dremio Cloud, supporting unlimited production use and infinite scale, with end-to-end security and SOC 2 Type 2 compliance. 

Find out more about Dremio Cloud here and sign up for it here.

Storage news ticker – March 3

By
Chris Mellor
-

…

File share outfit Box reported preliminary $233.4 million revenues for its Q4 ended January 31, up 17 per cent year-on-year. There was a net loss of $4.33 million, slightly lower than the year-ago $4.9 million. Remaining Performance Obligations (RPOs) went up 19 per cent to $1.1 billion. Full 2022 revenues were $874.3 million, up 13 per cent with a loss of $41.45 million, compared to a loss of $43.43 million in fiscal 2021. Next quarter’s revenues are expected to be between $233 million and $235 million, and 2023 revenues to be between $990 million and $996 million.

…

Research scientists have boosted DNA storage write speeds by increasing the number of letters in the DNA alphabet from the four (A, T, C and G) DNA base nucleotide sequences to seven by adding an extended molecular alphabet combining natural and chemically modified nucleotides. Data is encoded seven letters at a time and read through Oxford Nanopore sequencers. They say that, overall, the extended molecular alphabet may potentially offer a nearly twofold increase in storage density and potentially the same order of reduction in recording latency, thereby enabling new implementations of molecular recorders. But using current technologies, writing 100 bits of information requires nearly two hours. Writing 1MB will take nearly 10 hours. B&F thinks this it is commercially unusable, no matter how many years DNA storage is supposed to last and how dense it is (Library of Congress in a thimble and that sort of thing).

…

Microchip Technology announced the newest member of the Flashtec family, the PCIe 5 NVMe 4016 SSD controller with 16 high-speed programmable NAND Flash channels capable of up to 2400MT/sec, and the industry’s highest-performing controller of its kind. It addresses market demand for high-reliability, high-performance solid-state drives (SSDs) delivering greater than 14GB/sec throughput and 3 million IOPS. It has advanced Credit Management technology, supports NVMe v2.0, PCIe link encryption, and offers a stringent Quality of Service (QoS) required by cloud-focused datacentre applications. Solidigm thinks it is a good fit for TLC, QLC, and future NAND technologies, hinting at PLC. The NVMe 4016 is ready for Sapphire Rapids and its PCIe 5 support.

…

DDN-owned Tintri announced 42 per cent global revenue growth from 1H2021 to 2H2021, including a double-digit revenue increase from new customers. It is rebounding after a lukewarm few quarters and has a newly structured global executive sales team featuring Zachary Bertamini, VP Sales, Americas; Josh Marlar, VP Global Business Development;  Mark Walsh, VP Sales, EMEA; and Norimasa Ono, GM Sales, Japan. These are all internal promotions.

…

Intel still has plans for 3D XPoint

By
Chris Mellor
-

Contrary to the implications of recent news and presentations, Intel says it is not backing away from 3D XPoint, confirming to B&F that it plans to enable “our third-generation Optane Products with Sapphire Rapids.”

Since Micron withdrew from the 3D XPoint market, Pat Gelsinger took over as Intel’s CEO, and the company’s NAND non-volatile memory business is being sold to SK Hynix, questions have arisen about the future of Intel’s remaining non-volatile memory business.

This is the Optane SSD and Persistent Memory (PMem) operation, based on 3D XPoint technology, which was manufactured for Intel in a Micron fab at Lehi, Utah. Intel carried out XPoint research and development at its Rio Rancho plant in New Mexico. Second generation XPoint devices are currently being shipped, although third and fourth-generation devices were shown on a 2020 Intel roadmap, without dates ascribed.

However, over the past few months Intel has been silent about its Optane commitment and plans. From a product strategy point of view, Optane Persistent Memory, restricted to Xeon CPU connectivity,  has been seen as a way of helping Intel’s Xeon processors better compete with AMD’s server CPUs by effectively giving Xeons more memory, thus speeding applications.

Intel, under Gelsinger, is now looking to regain processor leadership over AMD and, with the aid of high-bandwidth memory, Xeon DRAM capacity limitations are being removed. That means the strategic necessity of having Optane PMem as a Xeon crutch is being lessened.

In February we learnt that Alper Ilkbahar, VP and GM for Intel’s Data Centre Memory and Storage Solutions unit, which includes the Optane business, is resigning. It also transpired that Intel had made more than a half-billion-dollar loss on its Optane 3D XPoint business in 2020.

Intel did not mention or promote Optane at its recent Investor Day event. Objective Analysis’s Jim Handy told us: ”It’s hard to tell how committed Intel is to 3D XPoint/Optane. I have seen no mention of it in management’s statements since Gelsinger took the helm. The silence is deafening… Read into this what you will. Intel is certainly not beating the Optane drum very hard.”

Ben Thompson’s Stratechery blog published an interview with Pat Gelsinger on 24 February, in which Gelsinger said: “I never want to be in memory, you see I’m doing everything I can to exit our memory businesses in that regard.”

Analyst Tom Coughlin wrote a Forbes article on 28 February, headlined “Is Intel Going To Drop Optane?”, where he commented: “It is likely that Intel will sell its remaining stock of current Optane products but not create new products for CXL and PCIe Gen 5 products.“

“Rumors abound that the company will not develop new Optane products,” he added, concluding: ”Intel hasn’t announced new Optane memory products for over a year and hasn’t created significant new 3D XPoint production since the Micron owned Lehi facility was shut down in 2021.  This once promising non-volatile memory technology may be in its last product generation.”

Sapphire Rapids slide from Intel Investor Day presentation

All this prompted us to ask Intel if it was committed to developing Optane and 3D XPoint. Ann Goldman of Intel’s Global Communications Group replied, saying: “We continue to work closely with customers and partners on memory and storage technologies, including enabling our third-generation Optane Products with Sapphire Rapids. In addition, we are enabling the ecosystem to be ready for key technologies such as memory tiering over CXL.”

That would place Optane PMem on the CXL bus for use as fast, non-volatile memory by applications running in servers connected to the CXL bus.

Intel is therefore not at the 3D XPoint exit point.

Bootnote

As a reminder, Optane is Intel’s brand name for devices built using 3D XPoint media. This is an implementation of Phase-Change Memory in which an electrical current is used to change the state of a chalcogenide glass material from crystalline to amorphous and back again. The two states have different resistance levels and these are used to signal binary ones and zeroes. Each state is persistent or non-volatile.

The XPoint media is fabricated in cells which are laid out in a 2-layer crosspoint array. Access time is faster than NAND flash but slower than DRAM, with writes taking about three times longer than reads.

Optane can be implemented as a memory bus-connected DIMM (PMem) or as an NVMe-connected SSD.

Chemical contamination at Kioxia, Western Digital NAND fabs cleaned up

By
Chris Mellor
-

The contamination of a chemical used to make 3D NAND that hurt output at Kioxia and Western Digital’s joint venture NAND fabs at Yokkaichi and Kitakami in Japan has been fixed.

WD said the plants’ contamination was resolved in late February, although it never clarified what the specific chemical was.

It said normal production has restarted and will ramp up in its third and fourth fiscal quarters, meaning the January-June period. Overall its flash output will be reduced by approximately seven exabytes. This is slightly more than the initial 6.5EB estimate and represents 29 per cent of WD’s 2022 Q2 output.

Kioxia’s equivalent statement was much blander: “As the manufacturing plants ramp back to full production, Kioxia anticipates that 3D flash memory BiCS FLASH shipments will be affected. The company will continue to make every effort to minimise customer impacts.”

According to Wells Fargo analyst Aaaron Rakers, WD and Kioxia take approximately 40 and 60 per cent shares of the NAND chip output from the two plants so we can expect Kioxia to lose about 11EB of output in the same January-June period, meaning an 18EB overall hit. 

WD reassessed revenue guidance for its third 2022 quarter, ending April 1, and now expects revenues to be $4.2bn-$4.4bn – $4.3bn at the mid-point. It had previously guided revenues for the quarter to be between $4.45bn and $4.65bn – $4.55bn at the mid-point.

This implies, Rakers says, that the output loss will cost WD $250m in its Q3 and we can expect a greater financial hit at Kioxia as it takes more of the fabs’ output. The loss should be lower in WD’s Q4 and Kioxia’s equivalent quarter as production will have ramped higher than the Q3 level.

But Rakers told subscribers he considers “NAND Flash price increases more than offsetting WD’s bit shipment impact – leaving us to consider increased F4Q22 estimates” for WD rather than lower ones. 

Kioxia and WD accounted for 33 per cent of the NAND market in the second 2020 quarter. Research house TrendForce agrees with Rakers, estimating NAND prices could rise by 5 to 10 per cent in the April-June quarter.

Pure Storage makes a profit for the first time as revenue rises 41 per cent

By
Chris Mellor
-

Pure Storage has made its first net profit, $14.9 million, reported for the fourth fiscal quarter of 2022.

The business reported revenues in the quarter ended February 6 of $708.6 million, 41 per cent up on a year ago – boosted by a “substantial” amount of FlashArray//C kit sold to Facebook parent Meta – and a first time net income of $14.9 million comparing to a net loss of $52.3 million for the same period last year.

Pure chairman and CEO Charlie Giancarlo summed it up: “By every measure, Pure had an outstanding quarter and fiscal year.” Full-year revenues were $2.2 billion, up 29.5 per cent, and there was a loss of $143.3 million, smaller than the prior year’s $282.1 million loss.

Growth rate has accelerated as Pure has at last moved into profit

Pure CFO Kevan Krysler said “We are thrilled to be capping off the year in a position of leadership and strength. The momentum we are experiencing is the year’s culmination of relentless focus on innovating for our customers. 

“Our performance in fiscal 2022 set new records for revenue, operating profit and cash flows. Our strong revenue growth this year benefited from a substantial sale of FlashArray//C to Meta, includes an extra week in the fiscal year and reflects an easier compare to fiscal 2021, which was impacted by COVID-related headwinds.”

Growth in the US outpaced other regions. Giancarlo said “The power of the IT environment in the US has been very strong,” partly because of demand for its file+object FlashBlade array. Krysler added, “The US and Canada, frankly, out-delivered for us.”

He said this about profitability: “Operating profit substantially expanded this year as a result of strong operating discipline and operating leverage. As we have previously highlighted, operating profits also benefited from slower-than-anticipated hiring, less travel and fewer physical marketing events due to the COVID environment.”

We suggested Pure could make a (GAAP) profit when discussing the prior quarter’s results because its growth rate looked so good at that point.

Pure has now passed the 10,000 customer mark, acquiring 470 new customers in the quarter, and the base includes over 50 per cent of US Fortune 500 companies. That means there is a lot of headroom for future growth – especially as disk arrays transition to all-flash arrays.

CTO Rob Lee alluded to this in the earnings call. “The transition from disk to flash is absolutely happening. We see it in the enterprise. We’re now starting to see it in the hyperscale environment. And we believe to a degree, this is not only happening, but it’s inevitable.”

Steepening growth (red) is only too apparent in this chart showing Pure’s quarterly revenues by fiscal year

Q4 subscription services revenue was up 42 per cent year-over-year at $216 million. Full-year subscription services revenue rose 37 per cent to $738.5 million. Krysler said “Our subscription net dollar retention or NDR at the end of the year exceeded 120 per cent compared to our long-term target of 115 per cent as a result of expansion growth from existing customers.”

Remaining performance obligations (RPO), meaning committed and non-cancellable future revenue, was over $1.4 billion, growing at 29 per cent.

Quarter-end finance summary

  • Gross margin – 67.2 per cent
  • Operating cash flow – $138.2M
  • Free cash flow – $117.2M
  • $1.2B in cash at quarter end
  • Total cash, cash equivalents, and investments – $1.41B

Giancarlo said in the earnings call “This past quarter, we again enjoyed strong growth, especially in the Americas, our largest theater and we grew across enterprise, commercial and public sectors. I’m especially pleased with our growth in both net new logos and subscription annual recurring revenue, indicating great progress against our long-term strategy.”

He pointed out “Pure’s proprietary hardware provides higher performance, reliability and longer lifetime while requiring less space, power and cooling than competitors’ commodity-based systems, competitors would need entirely new designs and years of new software development to replicate Pure’s advantages.”

Supply chain

What about supply chain issues, which have affected Dell, HPE and NetApp? “I feel confident that we have one of the most robust supply chains in the business. Pure has invested in strong relationships with our supply partners and flexible multi-source global operations over many years and we benefit from the architectural advantages of our integrated hardware designs. While there have been numerous supply chain challenges over the last year, and we are certainly not invulnerable to disruption in the future, we have navigated and managed them well and continue to meet our customers’ demands.”

Supply chain issues have become a new normal and Giancarlo doesn’t see any signs of improvement yet. “We had indicated last quarter that we thought it would bottom out in Q4. And it probably has, but it hasn’t really improved. It’s still bouncing along the bottom … we’re still waiting to see signs that we’re on the upswing.”

Hiring has been a plus point: “There has never been a time in my memory when so much great talent was available all at one time.”

The company has said it now has 4,200 employees.

Pure expects to see revenues of about $520 million next quarter, up 26 per cent year-on-year, with full 2023 revenues of around $2.6 billion – a 19 per cent rise on 2022.

Storage news ticker – 2 March

By
Chris Mellor
-

…

Datadobi, known for migration and data management ambitions, says the CBX agency has moved its data from an end-of-life NetApp system to a new one, using DobiMigrate migration software. It involved all of CBX’s client and internal business production data housed across two geographically dispersed locations in Minneapolis and New York City. Don Piraino, IT Manager, CBX, said: ”Typically, we use NetApp SnapMirror, but we knew it wasn’t the answer for this level of a migration. With DobiMigrate, we finished in weeks, rather than in months, with not even one hiccup.” 

…

FD Technologies has entered into a strategic partnership with Microsoft to expand the reach of its KX Insights streaming data analytics platform as more organisations look to modernise their data infrastructure for real-time decision making. KX Insights will be embedded natively on the Azure platform as a first class service, with KX generating revenue based on consumption. KX has also agreed to work with Microsoft to target new applications and services that will accelerate innovation and growth for organisations in the financial services sector. Adding streaming data capabilities is becoming a strong news trend – witness Druid, Imply, Ocient, Firebolt and Hazelcast.

…

ionir has bagged a US patent for its system and method for operating a storage system. It represents Ionir’s core architecture for its container native storage (CNS) technology: a Kubernetes-native, software defined, data services platform. Current storage systems and methods don’t enable users to store, retrieve or manipulate data in a storage system based on the logical context of the data. Typically, to store data in or retrieve from a storage system, users will need to specify a reference – such as a file name or physical location like a disc. Most CNS solutions still require complex backup, snapshot and restore. ionir is the only CNS technology that enables instant movement of data to/from any cluster, anywhere – and allows instant access to any point in time. 

…

Lightbits Labs has been assigned a patent (11,256,431) for “a storage system having a field programmable gate array.” The abstract states: “A field programmable gate array (FPGA), that includes a trusted FPGA logic, an untrusted FPGA logic and a monitor; wherein the monitor is configured to monitor the untrusted FPGA logic and prevent the untrusted FPGA logic from violating predefined constrains imposed on an operation of the untrusted FPGA logic; wherein the predefined constraints are stored in a memory region of the FPGA that is not accessible to the untrusted FPGA logic.”

Seagate and EVS, which supplies live video technology for broadcast and new media productions, announced a collaboration to provide mass capacity mobile storage and data transfer services for media and entertainment businesses. This will use Seagate’s Lyve Mobile edge storage and data transfer service.

…

IBM is partnering with Wanclouds to have its VPC+ Disaster Recovery as a Service and Multi-Cloud Migration sold on the IBM Cloud Marketplace. Companies can leverage VPC+ to orchestrate and manage on-demand cloud backups and multi-cloud migrations for cloud-native workloads and restore them when needed in the same or different regions within the IBM Cloud.  It can backup existing Kubernetes and OpenShift clusters on IBM Cloud and on-premises or edge Red Hat OpenShift clusters to IBM Cloud. VPC+ for IBM offers both a 30-day free trial (lite plan) and enterprise plan for its Disaster Recovery as a Service for IBM-managed Kubernetes Service & Red Hat OpenShift managed clusters. 

…

Cloud backup storage target Wasabi is now certified for use with Veritas NetBackup. It says users of Wasabi hot cloud storage and Veritas NetBackup benefit from more affordable, predictable pricing, simplified cloud storage and data management, and best-in-class security and recovery features to protect against ransomware and other threats, all at enterprise scale. 

…

Furiously fast filesystem producer WEKA has appointed Nilesh Patel as its first Chief Product Officer as it enters, it says, hyper-growth. Patel brings more than 30 years of experience to WEKA, having served in several product management and marketing leadership roles with Intel, NetApp, and Palo Alto Networks, among others.

Red Pro 20TB

…

Western Digital has expanded its Red Pro NAS disk drive range with a 20TB model having 9 x 2.2TB platters courtesy of its OptiNAND technology. This has a 64GB iNAND UFS flash drive embedded in the controller and enables more tracks on the platters. The new disk spins at 7.200rpm, has a 6Gbit/s SAS interface and a max sustained data transfer rate of 268BMB/sec; 4MB/sec slower than the 18TB Red Pro oddly. Endurance and workload parameters are the same as for the 18TB Red Pro.

…

Zadara (managed cloud edge services) has renewed its sponsorship of Sauber Motorsport, which manages and operates the Alfa Romeo F1 Team ORLEN. Zadara cloud services, which use the ultra-fast ISER interface with VMware compute, provide simultaneous access to real-time, data-driven insights in the driver’s seat, in the paddock, and in the team’s factory in Hinwil, Switzerland. The new Zadara-sponsored Alfa Romeo F1 Team ORLEN car, the C42, was revealed on February 27, along with team drivers Valtteri Bottas and Zhou Guanyu.

…

Index Engines’ CyberSense checks newer data first to speed up ransomware scans

By
Chris Mellor
-

Index Engines has claimed that its CyberSense data integrity scanner works up to 10 times faster with Dell’s Cyber Recovery vault because of operational changes.

Index Engines has software that catalogues unstructured data to create searchable indexes. These indices can be used for file tiering, eDiscovery, governance, and also ransomware detection with the CyberSense product.

Jim McGann, an Index Engines VP, said in a statement: “This new version of CyberSense breaks the performance barrier and allows clients to perform full content analytics on large data volumes daily.” 

CyberSense is integrated with the Dell PowerProtect Cyber Recovery vault. The intent of CyberSense scanning with Dell is to find the last good version of files and databases and so be ready to return a business to a pre-ransomware attack state. It provides full-content analytics and machine learning, which provides a 99.5 per cent level of confidence in finding signs of corruption due to ransomware.

Index Engines CyberSense video

The CyberSense machine learning analysis compares data as it changes over time to detect suspicious behaviour and cyber corruption. This full-content analytics approach is said to be more certain to detect corruption than just checking metadata.

Up until now, in an example where 400TB of virtual backups are replicated to a Cyber Recovery vault, full-content analytics would have required the 400TB of virtual images to be read and analysed daily, needing six CyberSense servers.

The new CyberSense versions, v7.8 and v7.9, prioritise recently modified data for analysis, allowing the most suspect data to be checked for integrity, and then scanning the balance of the data at a lower priority. The number of servers can be reduced to one, with an estimated 36TB of data read daily resulting in a up to 10X reduction in data read and analysed, using the change blocks of the client images.

It supports clients utilising Dell Networker, Avamar, and PowerProtect Data Manager as their backup products within a Cyber Recovery vault. This latest version reduces the Data Domain requirements and resources needed and, overall, provides significant cost savings for clients deploying Cyber Recovery with CyberSense, Index Engines says.

Get a datasheet here. The v7.8 and v7.9 versions of CyberSense are now available.

Quantum starts selling its custom hyperscaler tape library rigs

By
Chris Mellor
-

All those who said tape would soon be history can eat their words. Back in October Quantum revealed custom massive capacity modular scale-out tape libraries for hyperscaler customers. Now it has turned them into the i6H Scalar library product.

The i6H is for web-scale companies and enterprise customers needing petabytes to exabytes of tape storage capacity and is based on Quantum’s Redundant Array of Independent Libraries (RAIL) architecture, a RAID-like scheme providing scale, protection, and performance.

Bruno Hald, VP and GM, secondary storage for Quantum, said in a statement: “We now have seven hyperscale and web scale accounts globally that have collectively deployed over 35 exabytes of capacity in hundreds of Quantum tape systems around the world, including many Scalar i6H systems that are already deployed in some of the world’s largest data archives.”

He said the i6H was doing well with firms operating exabyte-scale private clouds.

Quantum’s moody shot of its i6H

Quantum has four products in its Scalar tape library line:

  • i3 with 25 to 400 tapes and 18PB compressed capacity with LTO-9 media
  • i6 with 50 to 800 tapes and 136PB of maximum compressed capacity
  • i6H with unrevealed tape slots and drive counts 
  • i6000 with 100 to 12,000 tapes and 540PB of compressed capacity in 21 racks

The i3, i6, and i6000 are delivered as individual chassis whereas the i6H is a scale-out library and delivered as a complete and fully assembled 48U rack. Each rack contains eight component chassis which look like Scalar i6 modules. Quantum is being coy about what is inside an i6H rack in terms of tape slots and drive counts, only saying it has best-in-class storage density.

But a rack of Scalar i6 chassis can hold 14.4PB uncompressed and 36PB of compressed data using LTO-9 tapes. Each 6U module supports up to three full-height LTO tape drives and 100 storage slots, for up to 4.5 PB (compressed) of storage per module. Therefore a rack of eight i6 chassis would total 24 tape drives and 800 cartridge slots.

On this basis a 21-rack i6H extravaganza configuration would hold 756PB of compressed data using LTO-9 tapes.

An i6H rack can be operational in less than 60 minutes from delivery at a customer site, Quantum says. All i6H components are customer-replaceable, even the robotic system. Quantum says the rack can accommodate additional servers to run software which can provide new ways to access tape for large-scale private cloud storage and archiving. 

The RAIL architecture can be used with Quantum’s ActiveScale Cold Storage software, which has two-dimensional erasure coding software that can write objects to multiple tape systems for more durable data archiving. The library includes anti-ransomware features, such as Scalar Ransom Block,  and comes with cloud-based analytics software.

The i6H has 80 Plus certified power supplies to reduce power and cooling costs. Tape integrity can be periodically checked with Scalar Extended Data Lifecycle Management (EDLM) policy-based tape scanning. The system includes monitoring and diagnostic functions and can be managed through a suite of RESTful web services.

Quantum has a Scalar i6H webpage discussing the product and the library is now generally available from Quantum.

Strong order growth at HPE but storage revenues drop 3 per cent

By
Chris Mellor
-

Revenues rose just 2 per cent in HPE’s first fiscal 2022 quarter, but order growth was up 20 per cent with as-a-service up 136 per cent. Supply constraints affecting virtually all IT system suppliers caused backlogs to build-up further and storage revenues slipped 3 per cent year-on-year.

HPE reported $6.96bn in revenues for the quarter ending 31 January 2022, compared to $6.83bn a year ago. There was a net profit of $513m, up 130 per cent year-on-year.

Antonio Neri, HPE’s president and CEO, said in the earnings call: “Hewlett Packard Enterprise delivered a solid performance. The quarter was characterised by robust customer demand and profitability… Our business pivot is… strengthening our gross and operating margins.”

The “pivot” refers to the GreenLake as-a-service business model. But there were “continued industry-wide supply constraints, which slowed our ability to convert orders” and contributed to a “record-breaking backlog,” Neri said. Both NetApp and Dell said the same dynamics had shaped their recent results.

Wells Fargo analyst Aaron Rakers told subscribers: “HPE noted that orders showed no signs of double ordering or noticeable cancelations and highlighted its belief that backlog hasn’t peaked.”

Financial summary

  • As-a-Service annualised revenue run-rate (ARR): $798m, up 23 per cent from the prior year period, and orders up 136 per cent
  • Gross margin: 33.7 per cent
  • Cash flow from operations: -$76m as end-of-year bonuses were paid to sales
  • Free cash flow: -$577m, reflecting normal seasonality and strategic working capital actions due to strong customer demand
  • Diluted earnings per share (EPS): $0.39 compared to year-ago $0.17

HPE splits its results across six segments, which all saw increased revenues except storage and financial services, as the table illustrates: 

Blocks & Files table

Storage

Storage revenues declined 3 per cent year-on-year to $1.192bn due, HPE said, to supply constraints on more complex owned-IP offerings – possibly its storage array ASIC hardware. Primera revenues grew 20 per cent annually based on constant currency.  We note that HPE did not say if Alletra 9000 revenues were up, the Alletra 9000 being the Primera follow-on array.

There was a storage segment operating profit margin of 14.5 per cent, down 0.51 per cent due to higher supply chain costs and an unfavourable shift in product mix, meaning HPE shipped more lower-margin products than it expected.

But storage product orders were up more than 15 per cent for the fourth consecutive quarter with a record backlog. So storage revenues were more affected by supply constraints than the Compute, Intelligent Edge, and HPC segments.

Dell reported flat storage earnings growth in its most recent quarter while NetApp revenues grew 9.8 per cent year-on-year. HPE is doing least well out of these three.

In the earnings call Neri said: “In Q1, nearly 10 per cent of computer storage orders were sold as a service.” The GreenLake services story is good news for HPE and it added more than 100 as-a-service customers in the quarter, taking the total past 1,350. 

Outlook

CFO Tarek Robbiati said: “We are off to a strong start delivering against our FY22 commitments with our third quarter in a row of more than 20 per cent year-over-year order growth bolstering our confidence for sustained revenue growth.” HPE is generating “better quality of earnings demonstrated by our improved gross margin despite ongoing supply chain constraints that enabled us to deliver Q1 EPS well above our outlook range and raise our outlook for the full year.” 

HPE forecasts diluted net EPS rather than revenue, with Q2 2022 EPS expected to be $0.18 to $0.26/share.

Smile – Fungible’s FunOS 4.0 brings DPU delight to entry-level systems

By
Chris Mellor
-

Adding server-offload cards to smaller numbers of servers has become possible with Fungible’s latest SW release.

Fungible is a Data Processing Unit (DPU) startup with two proprietary S1 and F1 chips that run its storage initiator cards and its scale-out  FS1600 NVMe array. The S1 and F1 DPUs run FunOS software on their data plane and it  features networking, storage, security, virtualisation, and analytics functionality.

The control plane runs Linux and contains agents that allow a cluster of F1 and S1 DPUs to be managed, controlled, and monitored by a set of REST APIs.

Fungible this week announced v4.0 FunOS to Blocks and Files, which adds support for smaller entry-level systems with reduced erasure coding and replica schemes. Chief Revenue Officer Brian McCloskey said: “One of the great things about the release of the 4.0 software stack is it provides erasure coding for smaller configurations. The sweet spot starter configuration for us prior to 4.0 was a six node cluster to enable 4 + 2 erasure coding. With the release of 4.0 it drives data efficiencies up and improves price per terabyte at smaller scale.” 

Erasure coding (EC) is an alternative to RAID and replication schemes, it has less overhead needed for the protection parity data with RAID and the copies needed with replicas. With erasure coding a data item is split into sub-sections with mathematical descriptive information (parity block) and each stored in a separate drive – or FS1600 node in Fungible’s scheme. If a storage node or drive fails or the data becomes corrupted, then the whole data item can be reconstructed from the other fragments with their parity blocks.

A 2+1 EC scheme divides each unit of data into two blocks with an added parity block, meaning 3 Fungible nodes are needed. A 4+2 scheme involves four data blocks and 2 parity blocks (6 nodes), while an 8+2 setup has 8 data blocks and 2 parity blocks (10 nodes).

The Fungible slide above illustrates that Fungible’s erasure coding scheme uses less node capacity overhead, than double or triple replicated alternatives protecting against single or double drive failure. The v4.0 release means Fungible kit can be used with 3-node clusters instead of needing 6 nodes.

McCloskey said: “The feature that we’re adding as part of FSC 4.0  is erasure coding 2 plus 1. We wanted to [have an] erasure coding level that lets a customer start small with our system, and then evolve and grow over time.”

Replication

Fungible also offers double replicated (RF2) volumes for extra data durability. The V4.0 SW brings in single replicated volumes which can be compressed. Fungible’s director for product management, Reese Loyd, said: “We’re able to layer compression and other data services on top of that. So the big one that we’re announcing as part of 4.0 is compression. You’ll be able to leverage single replica volumes that are both thin provisioned [and] compressed.”

Fungibles says it captures a lot of the raw performance of the NVMe drives in its FS1600 storage cluster, while also offering the compression data service. This can be used to extend application durability and to protect containerised workloads where speed is paramount.

Replication or erasure coding

Coldago analyst Philippe Nicolas told us: “Replication consumes lots of storage with x2 x3 the original size but it has very easy semantics. It is very good for small content, but when blocks become large it can create a penalty.”

“EC introduces striping plus protection (parities) and can accelerate throughput as multiple nodes collaborate, it also reduces the HW overhead compared to replication.” With Fungible’s dedicated DPU it’s pretty fast and very good for large blocks.

Replication requires fewer nodes and this makes it useful for disaster recovery.

He said applications such Cassandra come with built-in protection with 3 replicas at their level and don’t rely on sophisticated storage. Fungible announced RF1 to avoid having its own 3 copies at Cassandra level and redundancy at the cluster level.

“This comes from the fact that these applications were designed on commodity HW essentially with server-based storage (internal/DAS storage) in a shared-nothing model and they required some level of redundancy at the application level.”

Nicolas added: “RF1 … means that Fungible targets the kind of highly scalable databases used by hyperscalers and large corporations: Cassandra, MongoDB and they’re open source.”

“Keep in mind for such DB engines; they implement replication to avoid split brain problems and maintain data consistency.” With Fungible “they can still serve data with fewer nodes.”

Storage Initiators

Fungible announced its Storage Initiator cards in September last year and they need FunOS, but FunOS v4, with Storage Initiator support, is only being announced now.

Lloyd told us: “We announced the Storage Initiator in the fall and had functional internal code as well as using the functionality in our FDC product line. That is what allowed us to showcase the functionality at SDSC and others. [FunOS] 4.0 productises the feature fully for the Fungible Storage Cluster (FSC) and marks the general availability.”

Further: “The FSC 4.0 release represents the full product release including FunOS on the DPUs (both FAC cards and FS1600 nodes) and our composer (UI, API, telemetry, etc) software that runs on our composer nodes. This full package is FSC 4.0.”

FAC stands for Fungible Accelerator Cards and FDC is the Fungible Data Centre.

Futures

Lloyd said Fungible will “be able to layer additional data services on top of in the future.”

Fungible Storage Initiator cards.

The future may see Fungible’s storage initiator cards supporting non-Fungible storage targets. CEO Eric Hayes said: “We have no announcement on that yet. But we are very much supportive of supporting an open standard as NVMe over TCP for interoperability… Integration on the control plane, just like anything else, is where the work needs to be done. And we are entertaining this right now. We’re very supportive of these ideas.”

Another coming feature will be around composability. McCloskey said: “You’ll see an announcement from us in early April around our next step in the composable infrastructure market.”

Customers

Fungible’s boss said the company was making progress selling its products to customers. ”We have paying customers. Two of the world’s 10 biggest hyperscalars are customers. We have service provider customers. One of the world’s biggest retailers… is a customer.”

He also mentioned a very large telco, content delivery networks, and multiple US government entities as being Fungible customers. The company is hoping to drum up greater interest at large scale prospects in the aforementioned vertical markets, as well as with SaaS and managed service providers and larger enterprises.

It’s also looking for OEM opportunities.

Comment

The company is moving quickly, and it knows it has to do so. Intel is pushing its own version of the DPU as is NVIDIA with its BlueField (very) SmartNIC devices. These can run the vSphere hypervisor and thus provide features through added VMs. 

Intel and NVIDIA are established companies with strong channels and market presence. Fungible, as a startup has to punch well above its weight to make progress. It has a good base with its fast hardware and is building up its software features as rapidly as it can.

1...232233234...435Page 233 of 435

ABOUT US

Blocks & Files is a storage news, information and analysis site covering storage media, devices from drives through arrays to server-based storage, cloud storage, networking and protocols, data management, suppliers and standards. We publish news, views, opinions and analysis of storage industry events and developments, aiming to keep storage users, developers, consultants, marketeers, salespeople, support staff, analysts and everyone else involved in IT storage up to date with what’s going on.

FOLLOW US

© Situation Publishing, 2018-2024