Home Blog Page 130

Storage news ticker – June 23

By
Chris Mellor
-

Real-time data platform supplier Aerospike has unveiled a new graph database with native support for the Gremlin query language. Aerospike claims today’s graph databases fail to perform at scale with a sustainable cost model. Aerospike Graph delivers millisecond multi-hop graph queries at extreme throughput across trillions of vertices and edges. Benchmarks show a throughput of more than 100,000 queries per second with sub-5ms latency – on a fraction of the infrastructure.The Aerospike Database handles diverse workloads across three NoSQL data models – key value, document, and graph – in a single real-time data platform.

…

Data catalog and intelligence tech supplier Alation says its product is now available through Databricks Partner Connect. It says that data science, data analytics, and AI teams can now find, understand, and trust their data across both Databricks and non-Databricks data stores. This is accomplished by extending the data discovery, governance, and catalog capabilities built into Databricks and Unity Catalog across other enterprise data sources.

…

AMD EPYC Embedded Series processors are powering HPE Alletra Storage MP, HPE’s modular, multi-protocol storage product. This was first revealed in April this year. Now we learn that  Alletra Storage MP uses a range of EPYC Embedded processor models, with performance options ranging from 16 to 64 cores, and a thermal design power profile ranging from 155W to 225W.

EPYC Embedded Series processors are powering HPE Alletra Storage MP

…

A ransomware survey from Arcserve highlights potential vulnerabilities within local government and other public services. 36 percent of government IT departments do not have a documented disaster recovery plan. 24 percent of remote government workers are not equipped with backup and recovery solutions. 45 percent of government IT departments mistakenly believe it is not their responsibility to recover data and applications in public clouds.

…

AvePoint announced new functionality for Cloud Backup for Salesforce with FedRAMP (moderate) authorization, and the addition of the product to the AWS, Azure, and Salesforce AppExchange marketplaces.  Cloud Backup for Salesforce augments native Salesforce Data Recovery services by providing automatic, daily, and comprehensive backup of data and metadata and quick restores at the organization, object, record and field levels.

…

Open source real-time analytics, columnar database supplier ClickHouse has released ClickHouse Cloud on the Google Cloud Platform (GCP). General availability offers support across Asia, the European Union and the United States, and follows the December 2022 debut of ClickHouse Cloud on AWS. ClickHouse Cloud launched in December 2022 to provide a serverless, fully managed, turnkey platform for real-time analytics built on top of ClickHouse. Vimeo is a customer.

…

Catalogic’s CloudCasa business unit now supports the Red Hat OpenShift APIs for Data Protection Operator that installs open source Velero data protection on Red Hat OpenShift clusters for Kubernetes backups. The Red Hat OpenShift APIs for Data Protection Operator installs Velero and Red Hat OpenShift plugins for Velero to use for backup and restore operations. Red Hat is a contributor to Velero, along with VMware and CloudCasa, sponsoring many key features and maintaining Red Hat OpenShift APIs for Data Protection alignment with the upstream Velero project.

…

Real-time data connectivity supplier CData Software has announced that Salesforce Data Cloud customers will have access to select CData connectors that bring more data into Data Cloud from their SaaS, database and file-based sources. With CData Connectors, Salesforce customers will be able to streamline access to customer data across a wide collection of data sources and touchpoints.

…

Lakehouse supplier Databricks has introduced Lakehouse Apps, a way for developers to build native, secure applications for Databricks. It says customers will have access to a range of apps that run inside their Lakehouse instance, using their data, with Databricks’ security and governance capabilities. It also introduced new data sharing providers and AI model sharing capabilities to the Databricks Marketplace, which will be generally available at the Data + AI Summit, June 26-29.

…

Storage supplier DDN has been selected alongside Helmholtz Munich as a winner of the 2023 AI Breakthrough Awards in the “Best AI-based Solution for Life Sciences” category. DDN and Helmholtz Munich were recognized for their work in accelerating AI-driven discoveries that deliver concrete benefits to society and human health. Part of the Helmholtz Association, Germany’s largest research organization, Helmholtz Munich is one of 19 research centers that develops biomedical solutions and technologies. Helmholtz Munich has four fully populated SFA ES7990X systems that span a global namespace, an SFA NVMe ES400NVX system with GPU integration and EXAscaler parallel filesystem appliances.

…

Purpose-built backup appliance supplier ExaGrid has released v6.3 of its software with a security angle. Now only admins can delete a share, and all share deletes require a separate security officer’s approval. Two-factor authentication (2FA) is turned on by default. It can be turned off; a log is kept if that happens. RBAC roles are more secure as admins can only create/change/delete users and roles other than the security officer. Users with the admin and security officer roles cannot create/modify each other, and only those with the security officer role can delete other security officers (and there must always be at least one security officer identified).

…

Filecoin says its Filecoin Virtual Machine (FVM) is an execution environment for smart contracts allowing the first user programmability to the Filecoin blockchain. With FVM, developers can write and deploy custom code to run on the Filecoin blockchain, allowing them to connect, augment, and innovate around the building blocks of the Filecoin economy: storage, retrieval, and computation of content-addressed data at scale.  

The future holds making sure that FVM is easy to use and ensuring these onramps make sense for builders. Smaller storage deals, to allow building dApps and DAOs around pieces of individual, user-uploaded data. LLMs will play a central role in open-source generative AI models, all built as dataDAOs (DAO – Decentralized Autonomous Organisation) on top of FVM. There will be a decentralized aggregation standard, that will allow anybody to spin up “data dropboxes” for FVM.

..

HPE has added new capabilities to GreenLake for Backup and Recovery, which provides long-term retention, and protects data and workloads across customers’ hybrid clouds. It now protects on-premises and cloud databases managed by Microsoft SQL Server and Amazon Relational Database Service (RDS). GreenLake for Backup and Recovery also supports Amazon EC2 instances and EBS volumes.

…

Immutable and cryptographically verifiable enterprise-scale database supplier immudb has released immudb Vault. As an append-only database, immudb Vault tracks all data changes – preserving a complete and verifiable data change history – so data can only be appended, but never modified or deleted, ensuring the integrity of the data and the change history. Dennis Zimmer, CTO and co-founder of Codenotary, the primary contributor to the immudb project, said: “Before immudb Vault, businesses sometimes resorted to using ledger technologies and blockchains to share data with their counterparties, resulting in slow performance, complexity, high cost, and regulatory headaches.”

…

IBM is using AMD Pensando DPUs in its cloud. IBM says it has a 3 percent market share in the hybrid cloud market. Its niche is in the highly regulated public sector, financial services and airline customers including Canadian Imperial Bank of Commerce (CIBC), Delta Airlines and Uber. IBM is using the DPUs, with stateful security services, for virtual server instances and bare-metal systems. This allows IBM to unify provisioning across environments. It says this is pivotal in its strategy as it aims to continue double-digit revenue growth in its cloud business in 2023.

…

Reuters reports that NAND and DRAM supplier Micron will spend up to $825 million in a new chip assembly and test facility in Gujarat, India, its first factory in the country. With support from the Indian central government and from the state of Gujarat, the total investment in the facility will be $2.75 billion. Of that total, 50 percent will come from the Indian central government and 20 percent from the state of Gujarat. Construction is expected to begin in 2023 and the first phase of the project will be operational in late 2024. A second phase of the project is expected to start toward the second half of the decade. The two phases together will create up to 5,000 new direct Micron jobs

…

Following its acquisition by BMC, Model9 chief strategy officer Ed Ciliendo has left the company. His LinkedIn post says: “I am happy to announce that for the next few weeks I will assume the position of Chief Vacation Officer, managing a dynamic young team working on a broad range of KPIs from drastically increasing time spent in the sun to doubling our quota of ice cream consumption.”

…

Veeam’s 2023 Ransomware Trends Report found that 56 percent of companies risk reinfection during data restoration following a cyber-attack. This is because IT teams are often throwing precautions to the wind to minimize downtime, or are unsure of the proper safeguards to implement. ObjectFirst says the only way to ensure reinfection is impossible is through a combination of immutable storage to protect your data against attackers targeting backups as a vector for infection, and virus scanning/staged restores to ensure data that was already corrupted prior to being backed up does not make its way back. By themselves, these precautions are significantly less impactful at preventing reinfection, but with both in place companies can restore their data with confidence.

… 

Oracle‘s latest generation of its Exadata X10M platforms use 4th Gen AMD EPYC processors, delivering more capacity with support for higher levels of database consolidation than previous generations. Using EPYC CPUs allows for improved price performance, more storage and memory capacity, which enables more database consolidation and lowers costs for all database workloads. The EPYC-powered X10M  provides up to 3X higher transaction throughput and 3.6X faster analytic queries. It enables support for 50 percent higher memory capacity – enabling more databases to run on the same system.

…

Peak:AIO has doubled AI DataServer capacity by using Micron’s 9400 NVMe PCIe gen 4 SSDs instead of the 15TB drives used back in March. The then 367TB of capacity becomes 650TB in the same classic 2RU x 24-bay data server chassis. It provides up to 171GBps RAID6 bandwidth to a single GPU client using NVMe/TCP and ordinary, not parallel, NAS.

…

Decentralized storage startup Seal Storage, based in Toronto, has been selected as one of the 100 most-promising “Technology Pioneers’’ by the World Economic Forum. It says such Technology Pioneers are early-stage companies that lead in adopting new technologies and driving innovation, with the potential to have a significant impact on both the business landscape and society as a whole. Seal Storage was founded in 2021. Its software uses datacenters powered by renewable energy and operates on top of Filecoin DeStor’s open source technology to minimize its environmental impact. Seal offers zero egress fees to researchers allowing them to access their data around the globe.

What the WEF knows about decentralized storage can only be guessed at. Seal says it is poised to make meaningful contributions to the future of business and society through immutable, affordable and sustainable data storage – which gives a sense of the level of marketing fluff involved here.

…

SK hynix Beetle X31 storage
SK hynix Beetle X31

SK hynix has launched is first portable SSD, the Beetle X31 with 512GB and 1TB capacities. It has a DRAM buffer supporting the NAND which drives sequential write speed up to 1,050 MBps and reads to 1,000 MBps. Its heat management means it can transfer 500 GB of data at over 900 MBps. It is shaped, SK hynix says, like a beetle, with an aluminum case, weighs 53 grams and measures 74 x 46 x 14.8 mm. Host linkage is via USB C-to-C or C-to-A and there is a three-year warranty. 

…

SK hynix has received the automotive ASPICE Level 2 certification, the first time that a Korean semiconductor company wins the recognition. ASPICE (Automotive Software Process Improvement & Capability dEtermination) is a guideline for automotive software development that was introduced by European carmakers to evaluate reliability and capabilities of auto part suppliers. SK hynix will now aim for the ASPICE Level 3 certification with a more advanced software development process.

…

VAST Data is counting down for a 9:00am August 1 live-streamed Build Beyond event, saying: “The world deserves a single data engine for every workload to integrate Edge, Datacenter and Cloud. Data & Metadata. Unleashed Performance, Capacity & Insights! Build Beyond!” Register at buildbeyond.ai and watch a teaser video here. 

…

Vaughan Stewart, ex VP Technology Alliances at Pure Storage, has joined VAST Data as its VP of Systems Engineers.

…

AIOPs supplier Virtana has acquired cloud observability platform OpsCruise, the only purpose-built cloud-native, and Kubernetes observability platform. OpsCruise’s software helps predict performance degradation and pinpoint its cause. It has a deep understanding of Kubernetes and unique contextual AI/ML-based behavior profiling. OpsCruise software is integrated with leading open source monitoring tools and cloud providers for metrics, logs, traces, flows, and configuration data, freeing customers from proprietary alternatives.

NetApp names new chief commercial officer

By
Chris Mellor
-

Maxwell Long is the new chief commercial officer at NetApp, replacing Rick Scurfield who is departing two years after taking on the post.

Since joining NetApp in March 2021, Long has built and strengthened the North America Sales leadership team, and helped grow NetApp’s US Public Sector teams, US and Canadian Commercial. He comes into NetApp after being an SVP and chief customer officer at Adobe for just over four years. Before that he was a 12-year vet at Microsoft, finishing up as corporate VP for enterprise global delivery, leading a 10,000-strong team focused on delivering enterprise systems using Microsoft technologies.

Long will now target and expand NetApp’s enterprise, strategic, and commercial customer segments. He will drive global partner strategy and manage the execution of a deal life cycle, forecasting, territory management, and other processes and tools.

Max Long and Rick Scurfield, NetApp
Max Long and Rick Scurfield

Scurfield spent almost 21 years at NetApp, starting as a district manager for US DoD and intelligence agencies back in 2002. He then climbed the corporate ladder, becoming a regional director, then a VP and GM for OEM sales, SVP for the same, SVP/GM for Asia Pacific and Japan. A few more steps up the ladder took him up to the CCO spot.

He said on LinkedIn: “I am responsible for developing and implementing the strategic direction and global alignment across cross-functional organizations to accelerate NetApp’s growth while delivering best-in-class customer and partner satisfaction. To achieve this, I am leading a worldwide organization comprised of the Enterprise Strategy, Commercial Strategies, Channel and Partner Ecosystem, Customer Experience, Customer Success, Digital & Virtual Sales, the Global Strategic Technology Office, Go-to-Market Operations, Licensing, and NetApp Learning Services.”

In that role he also led NetApp’s Global Cloud Storage sales organization. NetApp’s public cloud revenues grew 25.8 percent year-on-year to $151 million in NetApp’s latest quarterly results. This was flat sequentially as public cloud revenues had grown 36 percent to $150 million in the prior quarter, up from $140 million in the quarter before that.

NetApp said: “As NetApp enters the next phase in its multi-year go-to-market modernization initiative, it continues to evolve the organization to better serve its customers and partners wherever they are in their respective transformation journeys. As part of that process, NetApp Chief Commercial Officer Rick Scurfield is leaving the company.

“The entire organization thanks Rick for his many contributions to NetApp over more than two decades, and the integral role he has played in driving NetApp’s growth and success over the years.”

DIY data science hits the Spot for MemVerge

By
Chris Mellor
-

Data scientists and developers can submit jobs to the AWS cloud in a DIY serverless fashion and MemVerge Memory Machine Cloud software will run the apps on Spot instances, says the vendor, adding that it will dynamically modify instance size as needed, with high-availability built in.

MemVerge supplies Big Memory computing software, combining DRAM and persistent memory DIMMs into a single clustered storage pool for applications to use with no code changes and run in-memory with less storage IO. It has developed so-called AppCapsules and in-memory snapshots which can can be used to deliver instant recovery for stateful in-memory apps.

MemVerge COO Jon Jiang said: “Memory Machine Cloud democratizes cloud computing by making it extremely simple to run apps in the cloud and by automating the previously complicated tasks of optimizing cost, performance, and availability.”

MemVerge says it set out to build software that enabled all applications, including stateful apps, to be serverless. The idea was that developers would never need to worry about configuration and optimization of servers or other public cloud resources, nor depend on another part of their organization or a supplier to do so.

AppCapsule technology allows MemVerge’s software to automatically move application runtimes across different machines and different clouds. It says this is the foundation for its Memory Machine Cloud software, making it possible for non-cloud experts to deploy, manage, and optimize apps in the cloud themselves.

Memory Machine Cloud eases cloud computing by making it simple to run apps in the cloud and by automating the somewhat complicated tasks of optimizing cost, performance, and availability. Non-expert users can have high-availability for non-fault-tolerant apps running on low-cost Spot instances, and have their workloads surf the AWS cloud to optimize cloud resource usage continuously during runtime.

AWS EC2 Spot Instances take advantage of unused EC2 capacity in the AWS cloud. Spot Instances are available at up to a 90 percent discount compared to normal on-demand prices, and can be used for stateless, fault-tolerant, or flexible applications such as big data, containerized workloads, CI/CD, web servers, high-performance computing (HPC), and test/dev workloads.

But Spot instances, as AWS spare capacity, can be reclaimed. Memory Machine Cloud deals with that through its SpotSurfer feature. This is a checkpointing and recovery service which gracefully recovers and restarts jobs when a Spot instance is reclaimed. This in turn makes it possible, MemVerge says, for long-running stateful workloads to run safely on Spot instances.

MemVerge customer Vince Pagano, a senior scientific programmer at TGen, said: “I was getting up to 80 percent batch failure rates with Spot EC2. Now with SpotSurfer we have already brought failure rates due to Spot reclaims to below 1 percent, and we are just getting started.”

There are three Memory Machine Cloud editions: Essentials which is free, Pro and Enterprise.

MemVerge Memory Machine Cloud tiers

Essentials includes includes the Float job automation and WaveWatcher resource and cost observability services. The Float automated job management service configures the cloud for an app owner, with jobs taking 1 minute or more to submit. The WaveWatcher rightsizing function profiles resource usage and identifies opportunities for optimization. It enables workloads to move automatically to smaller or larger instances during job runtime based on their real-time resource needs. 

Jiang said: “App users want to seize control of cloud computing and its cost without becoming a cloud expert or relying on another organization.” That reference indicated NetApp and its BlueXP Spot features to us.

NetApp announced Spot Ocean CD, a continuous delivery offering for Kubernetes, today. Ocean CD complements and extends Spot Ocean, optimizing delivery of cloud applications by automating deployment strategies across clusters and workloads. NetApp first announced a continuous delivery feature for Ocean, Ocean CD, in June 2021, when it automated app deployment and verification processes. It says the combination of Spot Ocean and Ocean CD extends its Kubernetes automation and optimization offerings to application delivery.

Memory Machine Cloud has achieved Spot-ready status with AWS and is integrated with the Jupyter Notebook and RStudio development environments and workflow managers such as Nextflow and Cromwell. AWS says this about partner Spot-ready status: “Amazon EC2 Spot Ready Partners enable customers to take advantage of Amazon EC2 Spot and unlock savings of up to 90 percent from on demand.”

MemVerge SimSurf

Users can get started with the no-charge Memory Machine Cloud Essentials. A SimSurf Simulator function on MemVerge’s website – user registration required – provides insight into how the overall scheme works. 

Commvault and VAST team up for faster restores

By
Chris Mellor
-

VAST Data and Commvault are developing a co-engineered system to store backups in less space and restore them faster.

Commvault’s data protection software is widely used by enterprises while VAST is growing quickly by supplying its single-tier all-flash storage for large deployments in enterprise datacenters. There is a natural market fit between the two suppliers who have an existing partnership. They are capitalizing on current ransomware concerns by saying they can generate and store backups safely, and restore them fast if a ransomware attack removes access to current files.

Commvault’s chief partner officer, Alan Atkinson, said: “Through this partnership and our upcoming joint offering, we’re helping solution and service providers and their customers to protect, secure, and ultimately recover their data at a much faster rate – keeping their business up and running, even in the wake of an attack.”

The two will produce a validated design that can create backups from public clouds, datacenters, SaaS apps, containers, and the edge, and store them on a VAST cluster. The system will have fast restores; “instant recovery” is the claim, from VAST’s infrastructure instead of slower restores from purpose-built backup appliances (PBBA).

B&F understands that at least one PBBA supplier has an all-flash system on its roadmap.

They say Commvault’s own compression and deduplication combines with VAST’s similarity-based data reduction to reduce the amount of data sent over network links and stored on the VAST system. This enables a system cost, they say, “equivalent to HDD-based backup appliances, with over one hundred times better restore performance.” 

No restore speed number was supplied. VAST currently claims it can “restore 50x faster than legacy PBBAs.” So, with Commvault’s help, it looks set to double that.

Commvault and VAST point out that VAST can run production and backup workloads on the same system, avoiding the need for a separate backup storage system, saving on cost, power and rackspace.

The VAST-Commvault system will take advantage of Commvault’s existing anti-threat measures, such as Metallic ThreatWise, multi-factor authentication, administrative workflows, Object Lock, and also VAST’s Indestructible Snapshots. The two say they will help customers to discover zero-day, unknown, and insider cyber threats.

Their validated design will have a detailed bill of materials and be offered on a consumption basis.

This VAST-Commvault engineered system idea follows on from a Pure Storage-Commvault reference architecture, with Pure’s FlashBlade as the backup store and Pure’s immutable snapshots safeguarding the data. That system delivers an up 270TB/hr restore speed; something for VAST to beat.

VAST is a backup target for Cohesity, Oracle (RMAN), Rubrik, Veeam, Veritas and HPE Zerto as well as Commvault. Commvault also has a partnership with Infinidat which has an SSA all-flash array in its product line. Polyamory rules in the backup software and target appliance world.

Commvault and VAST should have the validated design published and systems available from channel partners later this year. You can watch a video on LinkedIn if you want to see Alan Atkinson talking to VAST president Mike Wing about their two companies’ deal.

A last line of defense against ransomware

By
Rob Jaques
-
'recovery' key on keyboard
recovery key on keyboard

Sponsored Feature: The impact of cyberattacks around the world continues to escalate at an alarming rate, even after reaching “an all-time high” last year, new research warns. The latest IBM Cost of a Data Breach Report 2022 estimates that the average cost of a data breach hit $4.35 million in 2022, with some 83 percent of organizations found to have had more than one breach during the year. Business downtime can have devastating impacts on reputation and financial performance, and sixty percent of the surveyed organizations stated that they had been forced to hike the price of their services or products because of a data breach.

Ransomware cyber-crime remains “by far” the most common tactic employed by cyber criminals, with this type of attack launched approximately every 11 seconds according to calculations from Cybersecurity Ventures. These attacks can involve encrypting an organization’s backup data then demanding a fee for its decryption. Payment of these ransoms, David Bennett, Object First’s CEO, told El Reg, has been widespread among those desperate to get compromised businesses back up and running when they are unable, or don’t have sufficient time, to hunt down and restore previous versions of their data held on premises or in the cloud.

And the impact of the global ransomware problem is huge. IBM’s study estimates that, for critical infrastructure organizations (which the report defines as financial services, industrial, technology, energy, transportation, communication, healthcare, education and public sector industries), the cost of an attack is an eye-watering $4.82 million, considerably above the average. Some 28 percent of these were found to have been targeted with a destructive or ransomware attack, while 17 percent experienced a breach because of a business partner being compromised.

Bennett doesn’t mince his words when describing the potentially devastating impact of such ransomware attacks: “What can go wrong for your company if you are hit by a ransomware attack? The answer is everything. It doesn’t matter whether you’re a small company or a big company.”

“You need to consider all aspects of your business operations,” he continues. “How do you collect money from your customers, for example? How do you pay your employees? What happens if your ability to pay your employees or pay your suppliers is impacted? If you can’t collect cash, or you can’t pay cash, your business is out of business. And of course, there are all the other essential functions outside of just the financial operating ability of the company to consider.”

Ransomware scaling up and out

The scale of the danger is evidenced by the wide variety of global companies and organizations that have fallen victim to ransomware attacks over recent years. The WannaCry strain is reported to have affected telco Telefónica and other large Spanish companies for example, as well as the British National Health Service (NHS), FedEx, Deutsche Bahn, Honda and Renault, as well as the Russian Interior Ministry and the Russian telecom company MegaFon.

However, Object First’s Bennett notes that reported attacks represent just the tip of the ransomware iceberg given that so many organizations are typically very reluctant to admit being hit: “Unless they have a legal requirement, organizations are concerned about disclosing their vulnerability to cyber attacks,” he says. “The risk of reputational damage that can come from a data loss incident is enormous.”

Even the most robust traditional IT security measures – such as intrusion prevention, network protection, VPNs, DNS, and endpoint protection – cannot guarantee that mission crucial systems will remain beyond the reach of determined cyber criminals.

“It doesn’t matter what security systems you have put in place,” explains Bennett. “In the end, cyber attacks like ransomware are inevitable. The only way to deal with the situation is to be fully and properly prepared and know how to recover. The absolute minimum that any organization needs to have is an effective disaster recovery strategy.”

Three to one rule aids protection

Adhering to best practice may also mean abiding by the “three-to-one rule” which stipulates there are always three copies of data: one in production, one stored on different types of media, and one type which is immutable. This is exactly where, according to Bennett, Object First’s solutions can help by protecting object storage assets and making it impossible for ransomware-touting cybercriminals to encrypt an organization’s backup data.

Object First’s Ootbi platform has an unusual name, but is an abbreviation of “Out of the Box Immutability”, Bennett explains. It’s designed to be a ransomware-proof and immutable out-of-the-box solution which can deliver secure, simple and powerful backup storage for mid-enterprise organizations.

“We actually provide an immutable copy of our customers’ backups,” says Bennett. “This is important because historically, immutable copies were media like tape, or optical devices. And a lot of people offloaded their data to public cloud. You can have an immutable copy in the public cloud, be that AWS or Azure. But what happens if you must recover data that supports business critical systems from the cloud? It doesn’t happen very fast, and can you afford all that down time? If you have to rebuild all your systems it can take weeks or even months, and you must pay hefty egress fees.”

What Ootbi enables is a local immutable copy on-prem with a data separation from the storage control layer, which is the backup and recovery software, on Object First hardware. By separating the two and hosting them on-prem, customers can quickly recover, without high cloud egress fees.

Avoid paying the ransom

Bennett points out that, at a corporate level, Object First enjoys a long and close relationship with Veeam, a leading provider of modern data resiliency software and systems designed to deliver secure backup and fast, reliable recovery solutions. Ootbi is specifically designed for Veeam and Veeam only, which creates the experience of a near integrated storage appliance that combines all the necessary software, hardware and data management in one package. So, if a company already knows Veeam, they can implement an immutable storage solution without any previous experience.

Ootbi’s three-year subscription model with 24/7 support included means no surprises with fees in the long term. By delivering cheaper-than-cloud costs in a far more secure on-prem package, Object First argues that its customers can avoid paying the ransom and save money while doing it. Capacity and performance of the Ootbi locked-down Linux-based appliances scale linearly, supporting backup speeds up to 4.0 Gigabytes per second with up to half a petabyte of storage space.

This range of features made Object First’s Ootbi the perfect solution for Mirazon, a North America-based IT consulting specialist that helps customers with everything from basic helpdesk to complete infrastructure redesigns through managed services, consulting, and product sales.

“Mirazon required a solution that would not allow backup copies to be deleted or encrypted should they, or their customers, fall victim to a ransomware attack,” stated Brent Earls, chief technology officer, Mirazon.

The company also recognized the need to be able to properly secure not only its primary data, but also its backup data. As a result, it sought an effective ransomware-proof solution that would be simple to deploy and manage. Earls explains that cloud-based backups, while flexible, can be limited by bandwidth constraints: “Scale-out backup repositories are the only way of getting back-ups into the cloud, storage has either had to be re-architected to follow for smaller repositories to sync only critical data to the cloud, or the entire backup repository had to go all at once, which again, causes bandwidth issues,” he said.

“Implementing an on-premises device would solve the limited bandwidth issue while also eliminating the unpredictable and variable costs of the cloud.”

Choosing a solution that was both incredibly robust, but also simple to use and deploy was very important for Mirazon: “There was no existing product on the market that would solve the immutable issue that was also simple to deploy and operate which also fell within budget — until now. Object First has given us confidence that the solution will deliver everything it said it would.”

This focus on ease of deployment and use was also an overarching priority for Object First’s solutions, according to Bennett: “I‘ve been in the data storage industry for 20 years and the industry hasn’t really moved forward with the times. Historically, products are hard to use, hard to manage, hard to set up and you will need a storage management team to deal with managing backups. How many companies have a separate data storage team these days?”

“From opening the box to racking and stacking and getting it set up takes only 15 minutes. But in practice, most users have been able to do it in well under 15 minutes,” Bennett explains. “The hardest part is lifting the unit out of the box because it’s 180 pounds. But actually, implementing and getting up and running is super simple.”

Sponsored by Object First.

Micron doubles phone card flash speed

By
Chris Mellor
-

Using 232-layer NAND and the JEDEC UFS v4.0 standard, Micron has doubled its smartphone card speed compared to previous 176-layer product.

Update: Micron UFS card specificity for smartphone memory added, 23 June 2023. Mouser 1TB and 2TB Micron UFS 3.1 flash card capacity mistake removed, 24 June 2023.

SCSI-based UFS (Universal Flash Storage) flash cards are about the size of a fingernail. They can be used for embedded storage in smartphones and also for external storage on digital cameras and the like for data transfer and storage. The JEDEC UFS standard refers to both embedded memory storage and removable memory cards. Micron’s UFS 4.0 card is specifically for embedded memory storage in smartphones.

Mark Montierth, corporate VP and Micron’s Mobile Business Unit GM, said: “Micron’s latest mobile solution tightly weaves together our best-in-class UFS 4.0 technology, proprietary low-power controller, 232-layer NAND and highly configurable firmware architecture to deliver unmatched performance.” 

Micron UFS 4.0

Micron produced a 176-layer UFS card two years ago and its capacity range was 128, 256 and 512GB. (A Mouser reference to 1TB and 2TB Micron UFS 3.1 flash cards is wrong.) The latest card has a range of 256 and 512GB and 1TB, through using higher density NAND. 

It’s formatted as TLC (3bits/cell) with six planes to help parallelize access – four planes in 256GB model – and delivers up to 4,300MBps sequential read bandwidth, using Micron’s own controller. The prior product provided up to 1,890MBps so the speed increase is more than double.

But it’s puzzling that the older product could, at least according to Micron, download a two-hour 4K movie in 9.6 seconds, while the new one “allows users to quickly download two hours of 4K streaming content in less than 15 seconds, twice as fast as the prior generation.” We have asked Micron the obvious question – why is the new product, supposedly twice as fast, apparently five seconds slower?

The answer from Micron refers to increased download file sizes: “The download time for a two-hour 4K movie in 9.6 seconds provided in the UFS 3.1 release was based on a 14 GB file size. For our UFS 4.0 example in question, we based it on a 50 GB file size, which would be more standard for today’s flagship smartphones based on an average 7 MBps generated, equaling 50 GB for two hours. Based on the UFS 4.0 sequential write speed which is 4000 MBps (more than twice as fast than UFS 3.1’s 1500 MBps write speed), the download of the 50 GB file of 4K video would take ~13 seconds, and about 33 seconds for UFS 3.1, which is more than 2.5x faster.” That’s clear enough.

The new product’s sequential write speed is up to 4,000MBps and it is said to be 25 percent more power-efficient than the previous product. 

Micron is shipping samples of its UFS 4.0 storage solution to worldwide mobile manufacturers and chipset vendors. Volume production will begin later this year ready for the next wave of 5G and AI-enabled smartphones, Micron says.

SmartX releases IOMesh Kubernetes-native storage software

By
Chris Mellor
-

SmartX has released 1.0 of its IOMesh Kubernetes-native storage software.

Update. IOMesh is not open-source. 26 June 2023.

Beijing-based SmartX produces both hyperconverged infrastructure (HCI) software and IOMesh containerised storage software for Kubernetes environments. The preview version of IOMesh outperformed Ondat (bought by Akamai), Portworx (bought by Pure Storage) and Robin.io (bought by Rakuten), leaving SmartX as a stand-alone K8S storage software supplier.

A SmartX contact said IOMesh has “superior performance compared to Portworx and other peer products [and] is production-ready with enterprise features.”

IOMesh has gained Red Hat OpenShift certification and has been deployed by Alibaba Cloud and other customers. IOMesh v1.0 is the first commercial edition and complements the community edition by having better aspects of installation, deployment, storage, operations and maintenance. SmartX has offices in Palo Alto and Korea, so you are not limited to dealing with a company with a China-only physical presence.  

IOMesh cluster diagram

The software is said to be high-performance and features:

  • I/O localization – cross-node data reads can be minimized by storing a data replica on the local node and limiting read I/O operations to this node when possible, resulting in reduced network traffic and higher application performance.
  • Cold and hot data tiering – leverage the capacity and performance of different storage media types, maintaining a balance between cost and performance.
  • All-flash configuration – get lower latency and I/O variability, resulting in faster and more consistent application performance.
  • Local PV – create local PVs (persistent volumes) with block devices, providing better I/O performance than Kubernetes local PVs limited to directories.

High-availability and security functions include:

  • Multiple replica policy – place replicas between nodes based on local priority, optimal placement, and capacity balancing principles.
  • PV snapshot – take a snapshot of a volume in seconds to save and use historical data at any time.
  • Abnormal disk detection and isolation – automatically detect and isolate abnormal disks to reduce impact on system performance and operational burden.
  • Secure access – encrypt a PV using a StorageClass configured with Kubernetes Secret. Only users with the appropriate credentials can access the PV.
  • Intelligent data recovery – adaptively adjust data recovery or migration speed based on business loads to ensure business I/O.

Users can Install IOMesh Community Edition with a single line of code, and receive a free 30-day trial license upon installation. They can also later request a permanent free license. This installation requires an Intel x86_64 or a Kunpeng AArch64 Kubernetes cluster (3-node minimum). There is an IOMesh community on Slack for updates and community support.

IOMesh’s documentation provides an overview of its capabilities.

Xinnor, ScaleFlux squish the RAID rebuild

By
Chris Mellor
-

Xinnor says its software RAID with ScaleFlux compressing SSDs mitigates RAID rebuild impact, meaning less node-level redundancy with distributed data apps.

RAID is used to guard against drive failures when using multiple drives to provide the storage needed by applications. A distributed application, like a database, runs across multiple server nodes and these can need to be over-provisioned to cope with the extra internal traffic and degraded performance that come with failed device rebuild operations in RAID arrays. Xinnor provides xiRAID software with tunable IO parameters and SCakeFlux has computational SSDs with built-in compression technology to reduce device write amounts.

A Xinnor/ScaleFlux doc says: ”A key feature provided by xiRAID is the ability to control the rate of array reconstruction, enabling a predictable reconstruction workload sized to ameliorate contention with host IO. The The ScaleFlux CSD 3310 NVMe SSDs feature transparent inline compression that also lowers the impact of intra-RAID reconstruction IO.”

The two companies said they ran a test to see how a device failure and subsequent rebuild affected the IO rates of an Aerospike database accessing 38.5TB of storage capacity from 5 x 7.68TB ScaleFlux CSD 3310 PCIe gen 4 NVMe SSDs. The host server had a dual-socket Xeon Gold 6342 CPU with 48 cores and 512GB of DRAM. It ran Ubuntu 22.

In the setup Xinnor set its RAID 5 reconstruction priority to 10 percent of throughput to minimize the overall IO latency impact on the host. An Aerospike certification tool (ACT) was downloaded from GitHub and it generated a workload for the test and monitored the tail latency. The essential latency goal was that at most 5 percent of transactions exceed 1 millisecond.

The test involved four phases: normal operation, an SSD removal, degraded operation, rebuild when device returned to array, and normal operation again. A chart shows what happened to the IO patterns: 

Xinnor RAID volume throughput

There was no appreciable database IO impact at all, neither on reads nor on writes. The total read and write throughput are constant;at 5.4GBps and 1.7GBps, respectively. There is a slight decline in CPU utilization while the array reconstructs. The drive rebuild took about six hours, going by the chart.

The tail latency was similarly unaffected:

Xinnor tail latency

This all looks remarkable. RAID rebuilds normally are lengthy affairs, particularly with disk drives and particularly with high-capacity devices. ScaleFlux CSD 3000 series SSDs run the capacity gamut from 3,2TB through 3.84TB, 6.4TB, 7.68TB to 15.36TB. It would be interesting to see the test repeated with the 15.36TB drives, with the rebuild taking 12 hours assuming a linear progression. You could rebuild faster by setting the Xinnor Reconstruction priority differently.

The test shows that a Xinnor/ScaleFlux configuration can be used for low-latency applications. The two suppliers suggest there is an extension to distributed low-latency application scenarios. 

They say that “the increased node data reliability may make it feasible to reduce the amount of node-level redundancy. For example, deploying double replication in place of triple replication. Such an implementation would reduce server count and rack space, lower network utilization, and decrease the quantity of any required node-level licenses.”

InterSystems taps Volumez to provide storage

By
Chris Mellor
-
Amir Faintuch, Volumez
Amir Faintuch, Volumez

Block storage provisioning startup Volumez is rapidly expanding its capabilities, adding Azure cloud support to the existing AWS cloud and supplying storage for InterSystems’ IRIS cloud database.

Volumez provides a cloud-based control plane for provisioning NVMe block access storage for Linux-based, Kubernetes-organized applications running primarily in the AWS and Azure clouds but also on-premises. Users can specify capacity, performance, resilience and security needs with declarative statements. Volumez’s software composes or orchestrates the storage they need using Linux primitives and not the cloud supplier’s main block storage instances. GCP and other cloud support is coming as is an extension to providing file-based storage.

Scott Gnau, Head of Data Platforms at InterSystems, said: “The integration will allow us to deliver a fully-managed and high-performance database-as-a-service, seamlessly integrating AI functionalities within the IRIS kernel.”

InterSystems’ IRIS is a public cloud-based database management system for machine learning-enabled applications and supports real-time and analytics apps. It offers a unified data model that supports both structure and unstructured data, and has a set of APIs to interact with transactional data simultaneously in key-value, relational, object, document, and multidimensional databases. There is also an app development capability.

Basically Volumez, with its controller orchestration, can provide block storage for IRIS databases that is better than AWS’s own Elastic Block Storage (EBS).

Field CTO Brian Carmody told an IT Press Tour: “Volumez has 10x price/performance over EBS. We’re built on raw NVMe storage instance disk and EC2,  AWS’s raw NVMe instance storage (Nitro). There are no data services, no resilience but high raw IOPS … We guarantee 1.5 million IOPS, 300 microsec latency and 12GBps per volume. We support infinite volumes.

“We use off-the-shelf Linux to create file systems and mount points that outperform EBS. We use the cloud as bare metal. We are a control plane which runs in our cloud.”

Silk and Dell’s ScaleIO (now PowerFlex)  also use EC2 Instance storage but, he said, have proprietary technologies.

Carmody also said: “Silk is not cloud-native. It’s a controller architecture. CPUs on controllers will be the bottleneck in the future.” He included VAST Data’s DASE (Disaggregated Shared Everything) controllers in this statement. Volumez has a controller-less architecture.

John Blumenthal, Volumez
John Blumenthal

Volumez chief product officer John Blumenthal, who used to work at HPE, said: “We’re a composable data infrastructure company and not a storage company, though we run on storage media. I’m a traditionalist (controller storage) who’s lost his tradition.”

How does Volumez software build its storage? Carmody said: “We select a few slices of raw drives in a customer’s inventory to meet the need. We use the slices to set up NVMe instances and a data services storage stack in Linux kernel and present that mountpoint to the container or VM.”

Amir Faintuch, Volumez
Amir Faintuch

We suggested that AWS could do this. It invented the raw EC2 instance and it has clever Linux programmers as well. Volumez CEO Amir Faintuch said: “It [probably] will but we have a two-year start and patents.” There are more than 24 of these patents.

Blumenthal said: “CSPs make insane money from EBS and can’t afford to develop the stuff we have. They could do Volumez but economically couldn’t afford to do it.” He was referring to Clayton Christiansen’s theory idea of disruptive innovation, which was recently reviewed here.

Disruptive innovation

The Harvard Business Review article says: “As incumbent companies introduce higher-quality products or services (upper red line) to satisfy the high end of the market (where profitability is highest), they overshoot the needs of low-end customers and many mainstream customers. This leaves an opening for entrants to find footholds in the less-profitable segments that incumbents are neglecting. Entrants on a disruptive trajectory (lower red line) improve the performance of their offerings and move upmarket (where profitability is highest for them, too) and challenge the dominance of the incumbents.”

AWS is the incumbent and EBS the higher quality service/ Blumenthal and Faintuch are claiming Volumez, with its EC2-based storage, is the disrupter. Volumez and its backers are convinced that they have a persistent and defendable advantage in provisioning block storage public cloud apps that the CSPs cannot duplicate because it will destroy their highly profitable upmarket services, such as EBS.

Blumenthal added: “In the cloud the need for external-like storage is less. 80-90 percent of workloads benefit from DAS-like storage,” which is what Volumez supplies.

InterSystems head of product Luca Ravazzolo said he was “very pleased” with Volumez. Basically IRIS goes faster for less money with Volumez and the guaranteed performance is absolutely key.

Volumez says it’s working on replication between clouds and so providing a common data plane across the cloud providers. This is like Kubernetes for containers which is cross cloud.

Faintuch said: “We orchestrate at the OS level. Auto-provisioning could be coming and we orchestrate the infrastructure in the way that Polumi does and Terraform.”

Volumez is developing its software fast and needs to grow its user base and business before competitors try and muscle in on its act.

Starburst takes on Presto with distributed data lake source silo access

By
Chris Mellor
-

Starburst, the alternative Presto data lake analytics offering, is updating its fully managed Galaxy cloud service to support decentralized data lake activity, aiming to provide a single point of access for users to discover, govern, and analyze the data in and around their data lake. 

Presto is a Facebook (now Meta) 2012-originated open source project to provide datalake analytics using a distributed SQL query engine. Facebook contributed it to the Linux Foundation in 2019 which subsequently set up the Presto Foundation. The four Presto creators at Facebook left in 2018 and forked the Presto code to PrestoSQL. Facebook donated Presto to the Linux Foundation in 2019, which then set up the Presto Foundation. PrestoSQL was rebranded to Trino, and the forkers set up Starburst to sell Trino connectors and support. 

Justin Borgman, co-founder and CEO, Starburst, said: “Data teams are bogged down with complexity, often dealing with data silos within their organizations, forced to jump through multiple solutions just to understand what data they have and where it is located before they can even begin to think about its value.” 

Basic Presto architecture

Starburst wants to be the central gateway for the data lake and its associated silos. It’s announcing:

  • A fully managed platform with all the tools data engineers, analysts, and scientists need to activate the data in and around their data lake.
  • Multi-source access – Cross-cloud and cross-region federated access and governance making data from different sources accessible, allowing users to explore data before moving it into a data lake.
  • Connectivity – Great Lakes Connector provides connectivity to numerous data lakehouse file and table formats, including Apache Iceberg, Hive, Delta Lake, and more. 
  • Scalability – Warp Speed combines Starburst’s query engine with patented indexing technology for autonomous workload acceleration, increasing query performance.

Starburst says its query engine is deployed at PB-scale at the world’s largest internet companies.

Borgman said: “We believe Starburst Galaxy can act as an architectural centerpiece for modern data lakes that combines low cost commodity infrastructure with open formats and global federated access.”

Starburst has set up a Partner Connect portal to bring partner data source integrations to its customers. Partners for business intelligence (BI) and visualization include AWS QuickSight, GCP Looker, Metabase, Microsoft Power BI, Tableau Cloud, Thoughtspot, and Zing Data. Partners for data storage, prep and transformation are Tabular and dbt Cloud. 

These integrations are currently available to Galaxy users, with more to come. Starburst’s consulting and services partners like Accenture, Deloitte, Capgemini, and Slalom will be providing industry accelerators to help customers get started on Starburst Galaxy.

Generative AI and automation: accelerating the datacenter of the future

By
Nick Brackney, Dell Technologies
-

Commissioned: In the age of automation and generative AI (GenAI), it’s time to re-think what “datacenter” really means. For those who have become heavily invested in public cloud, the datacenter might not be the first place you think of when it comes to automation and GenAI, but these technologies are rapidly changing what is possible in all environments.

Ten or fifteen years ago, when businesses started bypassing IT by swiping credit cards and setting developers loose on cloud resources, the public cloud was absolutely the right move. In most large organizations, internal customers were often ignored, or their needs were not being fully met. They wanted flexibility, they craved scalability and they needed a low up-front cost to allow incubation projects to flourish.

If time stood still, perhaps the dire prognosticators of the datacenter’s end would have been right. I myself was quite the cloud evangelist before learning more about the other side of the fence. So why hasn’t this extinction-level event come to pass? Because the datacenter has adapted. Sure, there are “aaS” and subscription models now available on-premises; but the real stabilizing force has been automation.

Which brings us to the story of the day: GenAI and how it can augment automation in the datacenter to be an experience nearly on par with the public cloud. Before we get there though we need to look at the role automation and scripting have played in the datacenter. We’ll start by explaining some essentials, then we’ll unpack why automation and GenAI have changed what is possible on-premises.

Cloud operating model and infrastructure as code

Let’s start with the basics: the foundation of cloud was infrastructure as code and the idea of consuming IT as a Service. Your developers never had to talk to a storage admin, IT ops person, or the networking team to rapidly spin up an environment and get to work. This should be table stakes in 2023, and the good news it’s entirely possible to build it for yourself. Adopting this operational model means IT is leveraging policies and processes alongside automation to remove friction from the environment.

Visual representation of the end experience when you’ve automated a cloud operating model

Automation toolsets and telemetry data

Today there are many automation, management and telemetry/AIOps products available that provide unparalleled control and insights into datacenters. Data is the foundation AI and of managing a datacenter effectively. The control and visibility now in datacenters is often a superset of what can be achieved in the public cloud – although the hyperscalers have done a great job in that department as well. Given the cloud’s multitenant nature, cloud providers must obscure some of the operational knowledge to keep every customer secure. This results in architectural decisions that limit how some monitoring systems can be deployed and what data can be collected. One important are of focus is ensuring that you’re heavily integrating these solutions, embracing automation and infrastructure as code, measuring/monitoring everything and using a cohesive workflow for all your roles.

Visual representation of a common automation/management stack

The next wave of IT automation with GenAI

This brings us to the next evolution of the datacenter incorporating GenAI. Let me share a fun story about a past role where the client made the marketing consultant build an HCI deployment hands-on lab for physical and virtual infrastructure, and then didn’t provide any subject matter experts to help. If it’s not clear, that marketing consultant was me, and it was probably one of the most challenging projects I’ve ever worked on. I used code snippets and YouTube tutorials to get to the foundation of how to do such a task. I spent weeks assembling the puzzle, figuring out how each puzzle piece fit together. By some miracle I actually managed to get it right, even though I didn’t know much about coding. Anyway, here’s wonderwall… I mean here’s GenAI doing that.

GenAI is the Search Engine and code assembly machine we were looking for

Now mind you in my hands-on-lab, I was doing a lot more than just installing Windows Server, but there is no doubt in my mind if I asked it to provide the rest of that process, it could. What’s so important is that with the infrastructure-as-code mentality, and in new environments where developers may not be familiar with these types of calls or runbooks, GenAI is a new ally that can really help. Many people don’t realize access to common infrastructure scripts is prevalent – and oftentimes it’s written by the tech companies themselves. Both hardware and software vendors have large runbook repositories, sometimes it’s just a matter of finding them: enter GenAI. Another important consideration is that the infrastructure itself is intelligent and secure. These commands can be pushed out to thousands of servers for remote management purposes. This greatly lowers the bar on managing your environment.

GenAI and process building

One of my favorite customer engagement stories might sound a little long in the tooth – somewhat like those stories of being lost or unable to reach someone that are unfathomable to those who grew up with smartphones. We hear a ton of talk about containers, but when I broached this topic with one customer, he said, “I can’t even keep my VMware admins 18 months, what makes you think I could ever do containers?” This is something I’ve thought a lot about and it’s probably the biggest challenge with technology: if I don’t have the skillset, how could I possibly onboard it? Enter GenAI’s next incredible friction reducer: writing or finding documentation.

In just two prompts we have a routine and highly valuable process documented and ready to use

We’ve long had access to an incredible amount of information, however previously there’s been no ability to parse it all. This all changes with GenAI. Now, instead of navigating search and sifting through code repositories, a simple natural language query or prompt yields exactly the documentation needed. Instead of hours of looking for answers, extensive documentation is at your fingertips in minutes. This completely destroys any barriers to embracing technology. Imposter syndrome, skill gaps, and switching costs: you’re on notice.

Thousands of possibilities but AI Ops is next

I want to acknowledge the wealth of ways this technology can help us run a datacenter. Probably the next one to add significant value is AI Ops. That rich telemetry data can tell us a lot but also tends to have a signal-to-noise ratio problem. We’re simply generating too much data for human beings to analyze and comprehend it all. By pushing this data into GenAI and using natural language as an interface, we will extend insights to a broader audience and make it possible to ask questions we may never have thought of when looking at charts and raw data. The mean time to resolution will plummet when we use this kind of data. But there is one massive drawback, which brings us to our final point.

GenAI and automation change what’s possible, but we must use it carefully

Two of the major challenges with GenAI must be addressed. They are: Intellectual Property (IP) leakage and its ability to “hallucinate” or make things up. Let’s unpack each and determine how to embrace the technology without stumbling during implementation.

First, let’s discuss IP leakage. In any scenario where data is being sent to GenAI models that are delivered as a service, we risk leaking IP. Much like the early days of public cloud and open S3 buckets, early experimenters in their misuse or misunderstanding, created risk for their companies. The best way to counter this is to have a centralized IT strategy, insert them into your common workflows or development pipeline, and lastly prioritize building your own GenAI on-premises for highly sensitive data that cannot go to a AIaaS which is constantly learning off your data.

The other benefit of bringing a large language model (LLM) in house is you can also make it more precise and put guardrails on it. This makes the responses it generates more precise and in context of your own business. The guardrails can also stop some of the “hallucinating” i.e. when the GenAI is compelled to answer but provides inaccurate and/or made-up information to comply with the request. This is a common problem with GenAI. The reality is these tools are all still in their infancy. Just as most would work testing into their release pipeline, this too is an area where more rigor should be placed prior to pushing to production. I’m a big proponent of human in the loop, or human assisted machine learning, as a way to reduce mistakes with AI.

The future is automated

The datacenter is here to stay, but it can be radically transformed with GenAI and automation. These tools can augment our workflows and help IT Ops and developers achieve superhuman capabilities, but they are not a direct replacement for people. As you roll out your AI and automation strategies it’s important to think about what you’re trying to accomplish and what level of automation your organization is comfortable with. The future is bright and the ability to innovate anywhere is now a reality.

Learn how our Dell APEX portfolio helps organizations embrace a consistent cloud experience everywhere so they can adopt technologies like AI and accelerate innovation.

Brought to you by Dell Technologies.

Seagate bringing HAMR and online drive regeneration to Corvault

By
Chris Mellor
-

The Seagate Corvault product is getting online autonomous drive regeneration and a HAMR capacity leap.

Corvault is Seagate’s smart JBOD chassis with up to 106 3.5-inch drives, and its firmware running in VelosCT ASIC controllers. It currently runs with 20TB drives and has ADAPT erasure coding to protect the drive’s content and recover from failed drives. It also has ADR (Autonomous Drive Regeneration) to recover from a failed platter surface or head in a 10-platter drive. The drive is taken offline, reset to not use the failed head or platter surface, and then returned to service with reduced capacity.

Dave Thompson, Seagate EMEA System sales manager, told a briefing in Amsterdam: “Corvault will have 30+TB HAMR drives in the second half of this year.”

That accords with Seagate CFO Gianluca Romano’s comment at an analyst’s meeting that 32TB HAMR drives are coming. A Corvault with 20TB drive has a capacity of 2.12PB while one with 32TB drives gets capacity jumping up to 3.37PB, a 59 percent increase. In effect Corvault will increase its capacity by half a petabyte a year as HAMR drives go from 32TB to 36TB, on to 40TB and then in 2026 reach the 50PB level.

Seagate Corvault chassis
Corvault chassis with rear and side lids removed showing densely packed drives. The two ASIC controllers are at the rear left of the chassis in a smallish canister.

Other array suppliers have ADAPT-like drive protection capabilities but ADR is a proprietary Seagate technology. It is only available with Seagate’s 18TB and larger drives due to specific controllers and firmware being needed. In effect ADR removes the drive from the ADAPT pool and “remanufactures” the entire drive minus the failed surface. Seagate says ADAPT is designed for this scenario and can determine which erasure-coded chunks in Corvault’s drives are affected by the range of logical blocks now “missing” and thus need to be recovered.

Corvault is also getting an evolutionary advance in ADR, with the intent to have the drive stay online during its regeneration. Envisage a platter surface or head failing. Both events mean that the data on the platter in question cannot be accessed. Such enhanced ADR means that the recovery of the data in the failed section, using parity data on other drives, is 95 percent faster than an entire drive recovery. This is simply because one surface is 5 percent of the total capacity in a 10-platter drive. This shorter rebuild time effectively reduces the critical windows of system vulnerability when a second failure could happen.

Such a repair-in-place concept can extend a drive’s working life and reduce a customer’s e-waste amount because the drive is not going to landfill.

We should note that ADR can cope with more than one head/platter surface failure. This all helps to keep Corvault’s data availability time higher by reducing downtime. Seagate indicated to us that there were thousands of deployed Corvault systems, with one customer having 55 of  the chassis.

1...129130131...443Page 130 of 443

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