Mellanox has pumped an undisclosed sum of money into Excelero, its second announced investment in a storage startup in a week. The amount is likely to be small, as with its investment in WekaIO, announced yesterday.
Excelero provides NVMesh block storage accessed over an NVMe fabric and WekaIO provides file access which also uses an NVMe link between its front-end software and the target storage drives.
Both companies’ products are likely to be used where Mellanox’s Ethernet and InfiniBand switches are deployed. Excelero supports RoCE (Remote Direct Memory Access over Converged Ethernet) as well as InfiniBand.
In a scripted quote, Nimrod Gindi, head of investments at Mellanox Technologies, said: “Strategic partnerships with storage leaders such as Excelero and WekaIO are critical to develop the high-performance storage ecosystem and enable our customers to achieve efficient and scalable data processing and analytics capabilities to drive their businesses forward.”
Excelero CEO Lior Gal was equally effusive: “Mellanox’s low-latency networking and leading support for RDMA on both Ethernet and InfiniBand help us deliver the fastest distributed block storage solutions to our customers. We have partnered closely with Mellanox and welcome their investment in Excelero.”
Israeli-based Excelero was founded in 2014 and took in a $3.5m A-round and $12.5m B-round in 2015. In 2017 it received strategic investments from Micron, Qualcomm and one other unnamed business.
The company’s last reported funding round was in 2018 and involved Western Digital Capital. Total funding is $35m.
Update: Investment timing
Mellanox did not say when it invested in Excelero or the amount. We asked Excelero and Gal told Blocks & Files: “Mellanox’ investment with Excelero was an extension to Round B that was already announced last year. I can’t share the exact time or amount but it’s the fourth strategic investor we didn’t name last year.
“Mellanox, now with announcing us and WekaIO, decided to change strategy and be public about their bets in the future data center SDS technologies.”
Storage Moves Actifio, Cohesity and Commvault have all hired – or in Actifio’s case, lost – senior people as execs make career moves.
Jim Sullivan.
Actifio president Jim Sullivan has left to join NWN Corporation as the CEO and chairman of the IT services company. That leaves Actifio with a head of sales-size hole to fill.
His formal title at Actifio was President, Cloud Business Development, and his responsibilities included world-wide sales. Sullivan’s name has disappeared from Actifio’s leadership webpage and there is no named replacement.
Cohesity’s EMEA head hire
Cohesity has recruits Dirk Marichal to head up sales in EMEA . In his 20-year career Marichal has held EMEA sales VP-type positions at Avi Networks, Nutanix and Infoblox.
Dirk Marichal.
Two for Commvault
Commvault has hired Riccardo di Blasio as its Chief Revenue Officer, replacing Ron Miller, SVP w-w Sales who resigned in April 2019. The company recorded disappointing results in its latest quarter (Q4 fy2019); the first under new CEO Sanjay Mirchandani’s watch.
Riccardo di Blasio.
di Blasio’s resume includes global head of sales for VMware Cloud Platform Services at DXC Technology, CEO at Globetouch, COO at Cohesity, and SVP sales and marketing at VMware and EMC.
Blocks & Files notes Commvault has fairly regularly reported sales execution issues contributing to poor revenues, For example, here and here. It will be up to di Blasio to fix this problem.
Commvault has also appointed Sandra Hamilton as VP of Customer Success, responsible for Systems Engineering, Professional Services, and Support teams. The company says she will “drive end-to-end lifecycle engagement to support the increased adoption of Commvault’s subscription-based licensing and flexible pricing models.”
She joins from Puppet where she led the customer success team, and prior to that worked at exec-level in EMC and Accenture.
Nutanix grew faster than other HCI startups because it employed a core of ’10 eXer’ engineers and surrounded them with good people.
This was the central point made by Sunil Potti, Nutanix chief product and development officer, in an interview at the company’s .NEXT conference in Anaheim in May.
The other HCI startups were Atlantis, Maxta, Pivot3, Scale Computing, SimpliVity, and Springpath, with LeftHand Networks, acquired by HP in 2008, pioneering a virtual SAN idea. SimpliVity was bought by HPE and Cisco bought Springpath after initially OEMing its software in the HyperFlex offering. Atlantis and Maxta withdrew from the market.
VC-funded Pivot3 and Scale Computing are both continuing and growing, selling their HCI software through OEMS such as Lenovo and their own channel reselling partners.
Only Nutanx has made it to IPO status and it is the most successful HCI startup by far. Why is this?
In the beginning
Nutanix was started in 2009 by CEO and chairman Dheeraj Pandey, Chief Product Officer Ajeet Singh who left in 2012 to start up another business, and CTO Mohit Aron. Aron resigned in 2013 and subsequently started up Cohesity. Potti, the former head of Citrix’s netscaler business, joined Nutanix in early 2015.
He said Nutanix stood out from the pack because it had better engineering talent, with four or five core engineers who were far and away better than any other engineers in Silicon Valley.
Sunil Potti, Nutanix
Potti: “We call good engineers 10 eXers.” He said NetApp had similar class engineers when it was founded and Cisco too, in its router business.
Ultimately, he said, Nutanix executed better with 10X better people. The core design was very elegant and allowed mistakes to be made.
These engineers were bold. They built their own hypervisor, giving Nutanix more control over its foundation software. None of the other HCI startups did that. Dell, with VMware, has that level of hypervisor control as well.
The other HCI startups had engineers “maybe 50 per cent better than the average” but this was not good enough.
Potti, of course, spoke with the benefit of hindsight. Nutanix has succeeded, going to and beyond an IPO. So its engineers must have been good, have been 10 eXers, by definition. QED.
In Potti’s view such engineers can’t develop a stand-out product that leaves others behind. He thinks that the foundational product is like the basement of a building. You have to get the right core engineering talent to build it, because “it has to be strong. It can’t be brittle.”
Nutanix has carried on basing development on great engineers. According to Potti, “in every good service we build we base it around very good people and then surround them with other good people.” He makes it sound easy.
Part of the reason Nutanix has been able to do that because it has been a single product set company, with a mushrooming set of services expanding out from its core HCI foundation. Potti said incumbents such as Cisco, Dell EMC, HPE and NetApp, are multi-product set operations, “with internal competition for resources.”
“To do HCI well, you have to shed the old ways,” he argues. “Otherwise you get stuck in the messy middle…In big companies most of the competition is inside, not outside.”
There are resources invested in existing products which product line management s unwilling to give up.
Funding
Nutanix was also better funded than its HCI startup competitors, as a chart of funding rounds for HCI startups shows:
It accumulated $312.6m before its IPO in 2016. SimpliVity was in second place in funding terms, with $276m invested. It was bought by HPE for $630m in 2017.
Pivot3 is in third place in the total funding stakes, with $247m, followed by Scale Computing with $104m, and then Atlantis and Maxta with $35m each.
There is no guarantee that the more funding a startup gets the more likely it is to be successful. It wasn’t until 2014, five years after it was founded, that Nutanix’ funding soared. Ditto SimpliVity with its 2015 funding hike following six years of graft after it was started up.
Nutanix competitor tiers
Potti said the competitor Nutanix meets most in the field – its tier 1 competitor – is Dell EMC. Tier 2 embraces HPE and Cisco, fairly well-established HCI vendors. Tier 3 is NetApp.
We were interested in exploring Nutanix’s view of NetApp’s competitive status because NetApp has been a latecomer to HCI, tailing Cisco and HPE.
Can it get as established as Cisco and HPE?
Its HCI product includes a separate storage element, being a hybrid HCI product and therefore located in Potti’s messy middle ground.
Potti said NetApp’s customer base overlaps with Nutanix, so it’s not that defensible by NetApp in terms of keeping Nutanix out. Blocks & Files thinks it could get as established as Cisco and HPE but find it challenging to grow beyond that.
All-in-all, while expressing respect for the resources and talent in Cisco, HOE and NetApp, we came away convinced Nutanix is primarily focused on Dell Technologies now and AWS in the future as its main competitors.
WekaIO, the California analytics storage startup, has lined up a bunch of industry heavyweights to support its $31.7m C funding round.
They include HPE, Nvidia, Seagate, Western Digital Capital, Mellanox and Qualcomm, a previous investor.
In a press statement today, WekaIO said it will spend the money to bolster engineering, sales and marketing and to support overseas expansion,
To date the company has raised $66.7m at an unspecified valuation.
It has been a good few days for Weka which last week announced Mellanox as a customer.
Mellanox makes fast Ethernet and InfiniBand switches and will soon be a subsidiary of Nvidia, which agreed to buy the company in March 2019.
Mellanox switches are likely to carry WekaIO Matrix file data in high-performance computing (HPC) sites, with the data destined to be processed by Nvidia GPUs.
Mellanox uses Matrix in-house to gain faster access to data in its electronic design automation (EDA) workflow. It hopes to enable the design of more complex application specific integrated circuits (ASICS). The company said it bought Matrix because it was linearly scalable and supports InfiniBand.
Mellanox sells its switches into the HPC, artificial intelligence, enterprise data centres, cyber security, storage, and financial services markets, which are also served by WekaIO. It said the two companies share many joint customer deployments with deep learning workflows.
Mellanox said deep learning is computationally and data intensive, especially in the inference phase, and Matrix is well suited for this application. DL is also used in its own EDA workflows.
ConnectX 100Gb Ethernet networking from Mellanox was used in the WekaIO Matrix cluster design that broke performance records for the SPEC SFS 2014 and IO500 10 Node Challenge benchmarks.
Mimecast recorded yet another set of rising revenues at the end of its fiscal 2019.
Fourth quarter revenues for the email security and archiving vendor were up 26 per cent to $92.2m with a net loss of $1.9m.
Apart from slight Q3 to Q4 hiccups in fy’15 and 16 there’s a steadily rising revenue slope.
Free cash flow was $13.4m compared to $1.9 m a year ago. Cash and short-term investments at quarter end were $173.5m. The revenue retention rate was 111 per cent compared to 110 per cent a year ago, as existing customers renewed their subscriptions and bought additional services.
It gained 1,100 new customers in the quarter, taking its total to 34,400. The company seems able to recruit a thousand customers a quarter as regular as clockwork, with 4,000 gained in the whole fiscal year.
Full year revenues were $340.4m, 30 per cent higher than a year ago. The GAAP net loss was $7m, continuing the trend of trivial annual net losses we have recorded since its fiscal 2016 year.
Regular annual revenue rises and regular annual losses.
As long as the company continues to grow investors are more than happy to accept the losses which, anyway are decreasing as a proportion of the company’s revenues.
In the earnings call CEO, chairman and co-founder Peter Bauer said: “In our fourth quarter, our results exceeded our expectations for both revenue and profitability. Revenue of $92.2m grew at 26 per cent year-over-year as reported and 32 per cent in constant currency. We signed over 36-figure deals, a new high aided by the continued adoption of our multi-product bundles and traction with larger accounts.”
Mimecast’s customers are using more of its services as the cyber threat landscape becomes more populated. They used an average of 3.1 services in the quarter. It was 2.9 a year ago.
Co-founder and CTO Neil Murray resigned at the beginning of April. Bauer said: “He remains on the Board of the company, remains a very significant shareholder, very interested in the success of the business.”
Revenue in the next quarter is expected to be between $96.7m and $97.7m, $97.2m at the mid-point and a 23.9 per cent rise on the year-ago quarter. The rising revenue trend should continue unabated.
IBM is using embedded Everspin MRAM to increase the capacity of its proprietary flash drives used in its Storwize and FlashSystem arrays.
Everspin CEO Kevin Conley said during an earnings call this week that IBM is using MRAM in the Storwize V5000, adding to the existing Storwize V7000 and FlashSystem arrays .
IBM described its use of MRAM in its FlashCore modules (drives) at an MRAM Developer Day event in August last year. The company said Storwize arrays used system-wide protection measures to write the contents of DRAM to flash storage if power failed.
The arrays used proprietary format FlashCore modules and IBM wanted to change these to standard format U.2 (2-5-inch) flash drives. The drives use its FPGA controller which puts incoming writes into DRAM. To protect that data against power loss the choice was between super-capacitors or MRAM.
Super-capacitors are relatively large whereas MRAM is small and permits more space inside the drive for NAND. It is almost as fast to access as DRAM at 35ns and has unlimited endurance. However, it is only available in small capacities, with 256Mbit chips in production and 1Gbit varieties in sampling. 256Mbit chip are large enough for write caching needs and the drives have 128MB of MRAM for this purpose.
The U.2 bandwagon
In April this year IBM introduced its U.2 format FlashCore drives with 4.8TB, 9.6TB and 19.2TB capacity levels, using 64-layer TLC (3bits/cell) 3D NAND and the Everspin MRAM write caches.
FCM stands for FlashCore Module.
They are used for write cacheing plus journalling, logs, data buffering and streams in IBM’s Storwize V5100, V7000 and FlashSystem arrays.
The V5100 arrays have NVMe-connected FlashCore drives, using PCIe gen 4 x4 and support NVM-oF host access. The drives can configure the 4 PCIe lanes into a 2+2 scheme for dual-port access.
IBM capacity advantage at risk
The 19.2TB FlashCore module capacity is higher than many other SSD suppliers.
For example, Micron’s 9300 Pro tops out at 15.56TB using TLC flash, as does Western Digital’s Ultrastar DC SS530. However, Samsung has introduced a 30TB PM1643 SAS drive which blows IBM’s capacity advantage away – so long as you don’t need NVMe access. The same is true of Toshiba’s 30TB PM5, also a SAS interface drive.
Infinidat is trialling a software-defined all-flash array to sit inside Infinibox.
The high-end array maker developed the software in response to the requirement of Epic Systems, a prominent US healthcare software provider, that its health records application must use all-flash arrays.
Infinidat said other unnamed software vendors also specify all-flash arrays for their applications. That rules out Infinidat, which uses disk-based architecture with memory caching in its Infinibox arrays.
But the company has developed a cunning plan to get itself considered for certification for Epic Systems and other AFA-only specifiers. In an Infinidat CrowdChat last week CTO Brian Carmody said: “We’re announcing a beta feature called ‘Epic Compatibility Mode’ for such apps. It creates a software-defined solid state array within InfiniBox.”
The Infinibox OS ensures Epic data is stored only on SSDs in the array and not on disk. This can be managed via policies for specific datasets.
It will be a memory-cached all-flash array and should still run faster than ordinary all-flash arrays. Neat.
Infinidat’s Stanley Zaffos, SVP for product marketing, said in the Crowdchat that big software vendors “often dictate that their software only run on the fastest arrays to put themselves in the best possible performance light. Oracle, for example, recommends hosting their databases on RAID1 protected storage.”
Everspin, the niche MRAM developer, will need to raise new funds fast – or risk going bust. The loss-making company can self-fund at its current cash burn for another four quarters.
The company’s fundamental problem is that MRAM is faster than competing technologies but costs too much to manufacture. This makes it difficult to justify on price-performance grounds and MRAM take-up is slow. Can it overcome this?
Spinning class
Everspin last week posted $10m revenues for the first calendar 2019 quarter, a 33 per cent fall on last year’s $14.9m. Net loss was $4.3m (-$3.5m Q1,2018). It expects revenues of $9.6m to $10m in the next quarter and a similar loss. Cash and cash equivalents at the first quarter end were $18.5 million; compared with $23.4m a year ago.
Revenues come mainly from technology IP licensing and royalty payments by MRAM device manufacturers. It attributed the revenue fall, mainly to a one time IP licensing fee realised in Q1 2018.
Everspin has had $80.3m in funding and IPOed in 2016. The company raised $24.5m in a secondary offering in February 2018. It also obtained a $12m loan and security agreement with Silicon Valley Bank in 2017. The company is taking a hatchet to R&D costs by teaming up with GlobalFoundries through a joint development agreement. R&D costs fell in Q1 from $6.5m to $4m. But Blocks & Files expects the company will need another funding raise in the next 12 months.
Another spin
Our thinking is confirmed by Everspin’s 10Q SEC filing filed last week which acknowledges doubts over ability to continue trading;
These conditions, and our cash balances, raise doubt about the Company’s ability to continue as a going concern for at least a year after the issuance date of the accompanying condensed financial statements.
The Company plans to address these conditions by reductions in spending, improvements in product margins, increasing revenue, raising capital through equity or debt financings, licensing of its Intellectual Property, or a combination thereof. There is no assurance, however, that the Company will be successful in these plans.
Dire straits ahead
Management bleakly states that Everspin, if unable to raise additional capital or fund itself from its own operations “may be forced to cease operations altogether, file for bankruptcy, or undertake any combination of the foregoing. In such event, our stockholders may lose their entire investment in our company.”
There is a concern over its ability to fund itself as “our per-bit cost to produce our product is currently higher than competing technologies. As a result, our ability to capture market share and generate sufficient revenue to transition to profitability and generate consistent positive cash flows is uncertain.”
The company has identified and is fixing an internal weakness in financial reporting which, if not fixed, “could cause investors to lose confidence in our financial information or cause our stock price to decline”.
Customers on the horizon
In the earnings call last week, CEO Kevin Conley said: “IBM, our largest STT-MRAM customer, announced in April that the NVMe flash core module using our 256-megabit part is now supported in its third enterprise storage array family.” That is the Storwise V5000 product family.
Everspin MRAM is being used in IBM’s FlashSystem 9100 and Storwise V7000 products.
Conley also said three products from SSD controller companies designed-in support for its MRAM, and qualification was proceeding with one or more to be completed this year. These controllers would then have to be qualified in turn by the controller companies’ own customers.
Everspin secured two design-wins for its 28nm 1Gbit STT-MRAM product during the quarter and appointed Troy Winslow as Global Sales VP. Get selling Troy, before the walls come tumbling down.
Larger enterprises are storing multiple petabytes of data already and many are heading towards storing exabytes. This presents significant economic and management challenges, according to Infinidat, the high-end storage array maker.
‘Traditional hardware-based storage arrays are expensive, hard to manage, and orders of magnitude too small for the coming data age. They must, therefore, evolve into something new: software-defined on-premises enterprise storage clouds,” Infinidat proclaimed in a white paper published this month.
“The next generation of applications cannot be built on traditional enterprise infrastructure, because the economics and user experience required in the multi-petabyte enterprise are so radically different from what was offered in the past by traditional IT in the terabyte era.”
According to Infinidat the solution is to build huge scale-out capability for storage, moving away from single storage arrays towards a large cluster or private cloud of arrays.
To accomplish this, Infinidat is developing the Elastic Data Fabric, a concept based on ‘Availability Zones’ and built atop Infinibox array clusters.
Elastic Data Fabric will combine dozens to hundreds of storage systems across multiple on-premises data centres and data in public clouds into a single, global storage system scalable up to multiple exabytes.
It is based on Infinidat’s existing arrays and software and the company will deliver elements of the Availability Zone framework in mid-2020. AZs pools storage with seamless data mobility inside a data centre and the public cloud, and offer near infinite scalability, according to Infinidat.
Public Cloud influence
According to Infinidat petabyte-scale computing “on premises is lower cost than in the public cloud,” but customers need public cloud attributes when handling such large amounts of storage. In its white paper, the company lists the user requirements:
Fast, reliable, virtualised data storage on demand that is API-driven and seemingly infinite in supply. Think in terms of regions and availability zones, not arrays and filers
Storage services must be tightly integrated with compute tiers such as VMware and Kubernetes
Data must be secured 24×7
Customer business units should not be subjected to change control procedures and other hindrances related to data migration activities executed by the cloud operator.
Infinidat aims to fulfil this brief, courtesy of the InfiniVerse management tool.
EDF control plane
InfiniVerse acts as the control and data plane for Elastic Data Fabric. The software clusters up to 100 InfiniBoxes per data centre into a scale-out storage cloud.
Individual arrays, delivered in racks, are invisible to end-users in an availability zone. All storage objects including pools, volumes, and file systems are provisioned at the AZ level. AZs support heterogeneous InfiniBox configurations, such as capacity and performance-optimised racks, and multiple generations of InfiniBox hardware.
Data placement and mobility decisions within the cluster can be executed manually or by an AI agent that monitors and automatically responds to quality of service policy compliance. This is all controlled and monitored by InfiniVerse’s UI and REST API.
An Infinidat EDF diagram shows Elastic Data Fabric embracing multiple geographically separate data centres.
AZs are defined as operational within data centres, and as yet there is no management facility to handle multiple AZs across data centres. It seems reasonable to infer that Infinidat will add this capability in due course.
EDF components
Six Elastic Data Fabric components are available today:
InfiniBox arrays
Neutrix Cloud – Infiniboxes in a data centre connected directly to public cloud operators
InfiniBox FLX – InfiniBox as a service
InfiniVerse SaaS storage monitoring and analytics and non-disruptive box-to-box data mobility
InfiniGuard data protection
InfiniGuard FLX – InfiniGuard as a service
Two components are planned for mid-2020:
InfiniVerse storage AZs
InfiniVerse transparent data mobility (TDM) with no server or application awareness
Data mobility
Elastic Data Fabric scales as data sets grow or temporary workloads are created, and shrinks as they retire, morph, or move closer to the applications and users over time, Infinidat said.
There is no need for data migration change control procedures due to hardware refresh.
Non-disruptive data mobility (NDM) is live today and enables workload migration of workload from one InfiniBox to another without application downtime. It uses InfiniBox Active/Active replication and Host Power Tools to provide a scriptable data migration process between any two InfiniBox systems within metro distance.
Transparent data mobility (TDM) will provide workload mobility from rack to rack within an Availability Zone, with no downtime and no involvement or awareness at server level. It will use a distributed peer-to-peer function on the target port drivers and does not need in-band virtualization appliances.
Exabyte-scale physical SAN versus hyperconverged system
Infinidat provides shared external storage arrays accessed by applications running in connected servers. This contrasts with the hyperconverged approach which combines servers and storage in connected nodes running as a virtual SAN. There can be compute and storage nodes in such a scheme but the virtual SAN concept is still used.
We note that an InfiniBox F6300 has up to 4PB of usable capacity and up to 10PB of effective capacity, after data reduction. A hundred systems means an AZ could have up to 400PB usable and 1,000PB effective capacity.
We think storing 400PB usable capacity in a hyperconverged system would be a huge challenge. With Nutanix, for example, this requires 5,000 NX-6035C-G5 storage-only nodes, each with 80TB raw capacity. And you still need compute nodes, resulting in a substantial system management task to handle multiple thousands of nodes.
That is a different order of complexity compared with handling 100 Infinibox arrays as an entity.
Here is a snapshot of disk drive supplier market shares based on capacity for first 2019 quarter.
Western Digital – 45.8 per cent
Seagate – 40.9 per cent
Toshiba – 13.3 per cent
In the enterprise nearline drive market the supplier positions were:
Western Digital – 55.7 per cent
Seagate – 35.2 per cent (its lowest share since 2014)
Toshiba – 9.1 per cent
Aaron Rakers, a Wells Fargo senior analyst, suggested hyperscale buyers are responsible for a jump in Toshiba’s share from 5.9 per cent a year ago.
He has supplied a chart showing trends in disk drive revenue, unit and capacity shipments:
And it’s a hard drive gonna fall
Revenues and units are declining as SSDs replace disk drives for mission-critical enterprise data storage and in notebooks and PCs. However, enterprises are storing more secondary data in nearline drives and that is driving capacity increases.
The chart above shows that revenue per HDD market segment is going downwards except for nearline drives – although that is a bit of a roller-coaster recently.
Some enterprise storage snippets for your delectation, filleted for ease of digestion. Let’s start with Dropbox.
Dropbox puts cold data in a Magic Pocket
Dropbox has added cold storage to Magic Pocket, the exabyte-scale storage infrastructure developed in-house as a replacement for AWS services.
The new storage tier is optimised for less frequently accessed data and is not marketed as a standalone service.
The cold storage tier uses the same Magic Pocket storage infrastructure, including SMR drives, as main or warm data “but saves costs through innovatively reducing disk usage by 25%, without compromising durability or availability”, Dropbox wrote in a blog post. “The end experience for users is almost indistinguishable between the two tiers. “
This accomplish this, the company has devised a way of writing files into storage blocks with less replication across geographic regions. Instead of fully replicating files regionally, the new method divides a block of storage into fragments and stripes these across three regions. Read requests are sent to the three regions and the fastest response from two regions is used to build the block for the user.
This takes a little longer than the time needed for warm data.
Data is tiered automatically – it is not under user control. Dropbox must log file activity and transfer little-accessed data into a cold storage state.
Amazon has a separate Glacier cold data storage service and infrastructure, and Azure and Google offer similar services..
Micron costs for buying out Intel JV
Micron has confirmed it will pay $1.3bn to $1.5bn to buy out Intel’s interest in their joint IMFT venture and its flash and 3D XPoint fab in Lehi. The JV will end on October 31, 2019. The price to pay and date of closure is detailed in an SEC regulatory filing by Micron.
Micron will continue to use the Lehi plant and the two suppliers will then compete and separately sell and market XPoint devices and, potentially, media. This should widen the XPoint market and hopefully bring prices down.
SK Hynix dances the quad-level cell 96-layer dance
SK Hynix is shipping samples of 96-layer 3D NAND chips to SSD manufacturers.
The Korean semiconductor giant currently ships 72-layer 3D NAND chips and the new product will keep it in the race with 3D NAND competitors which already have 96- or 90-99 layer-tech (Samsung is coy about the precise layer count).
The chips use CTF (Charge Trap Flash) technology and are formatted as quad-level cell (QLC) flash with 4 bits/cell. SK Hynix said CTF “has reduced the area to less than 90 per cent of the existing 3D-based QLC products”.
These SK Hynix chips have 4 planes for data access, providing 64KB bandwidth. As well as shipping samples to SSD suppliers, SK Hynix is developing its own QLC NAND controller and code.
Han Joo Na, VP and Head of SK Hynix’s NAND development strategy office, said: “We plan to launch our own QLC-based SSD from next year when demand for enterprise QLC NAND is expected to become meaningful.” The company is looking at the enterprise SSD (eSSD) market, with 16TB or larger drives replacing 3.5-inch disk drives.
SK Hynix cited IDC forecasts that QLC will increase from three per cent in the NAND flash market in 2019 to 22 per cent in 2023. The eSSD market will grow at a compound annual average growth rate of 47.9 per cent in GB unit terms, according to IDC, which, SK Hynix said, means QLC flash will rapidly replace disk drives.
SwiftStack goes Splunk
Object storage software supplier SwiftStack has announced a software appliance with “1click Application Profiles…to configure a ready-to-use storage system tailored for the specific application”.
The company’s first application profile is for Splunk Enterprise with SmartStore and it has made the SwiftStack Software Appliance nitially available for Cisco’s UCS S3260 storage server, with Cisco offering a 3-chassis UCS S3260 configuration with 672TB of usable storage capacity.
More application profiles and supported storage servers are on their way, the company said.
.NEXT 2019 Nutanix’s annual conference was in full swing at Anaheim, California this week, and the news announcements from company and partners came thick and fast.
Here is a drawerful of news shorts for your perusal.
At the event Commvault and HYCU introduced new data protection capabilities for Nutanix and Fujitsu added a server hardware platform to run to Nutanix hyperconverged infrastructure (HCI) software. Nutanix has introduced wider cloud support to Frame, Xi Leap and Beam.
Commvault & Nutanix Files
Commvault said Nutanix Files 3.5 is an enterprise-grade scale-out file storage system that leverages the cloud-like flexibility of Nutanix HCI architecture.
Brian Brockway, chief technology officer, Commvault, said: “Nutanix Files expands the HCI value proposition into a top-tier enterprise data solution.”
Commvault’s Complete Backup and Recovery data protection product is integrated with Files 3.5 and features include:
Snapshot-based backup copy protection for CIFS (common internet file system – Windows file shares) and NFS (network file shares/Linux/Unix file shares)
Change File Tracking (CFT) API integration
Auto-discovery of Nutanix Files data share
Granular recovery options—files, folders, versions for example
Commvault Storage Policy integration, allowing backup data to be stored onsite, offsite, or in the cloud
Recovery to the original Nutanix Files share or any compatible file location addressable by Commvault
Support for analytics and data governance use-cases featuring Commvault Activate
Simplified configuration and operations
SLA-based management, compliance and reporting
Integrated support for Nutanix Files 3.5 is included in the current release of Commvault Complete Backup and Recovery.
HYCU for Nutanix v4.0
HYCU has added backup support for NFS Share on Nutanix files.
HYCU uses Nutanix Snapshots for backup and recovery. HYCU for Nutanix v4.0 supports backup of VMs using NearSync volume snapshots, running on AHV or ESX. Application data recovery can selectively restore only the disks needed to get the right application data.
Customers can implement different backup policies for different parts of their file shares.
HYCU has added support for applications and data running on Windows Physical Servers. It has also added a built-in option to turn on call-home functionality. That enables HYCU to get working on a problem fix and notify the customers of any major issue before they might have noticed it.
A single deployment of HYCU can manage multiple Nutanix and VMWare clusters, with a consolidated view.
HYCU for Nutanix v4.0 also has validated support for Nutanix Buckets. It should be generally available within 30 days.
Nutanix Frame, Xi Leap and Beam
Nutanix has extended the Xi Frame desktop-as-a-service to the private cloud, enabling the delivery of apps and desktops in hybrid cloud environments. Xi Frame already supports AWS and Azure, with access using any browser and any device.
When running on AHV, customers can extend desktop delivery to their Nutanix private cloud, integrating virtual desktop infrastructure (VDI) services with the Nutanix Enterprise Cloud platform. Nutanix said customers can provision thousands of virtual desktops in minutes.
The company also announced new functionality for its cloud-based disaster recovery (DR) service, Xi Leap. Customers can configured policies to automate the DR workflow directly within the Prism Console and restoration can take place in minutes, Nutanix said.
Xi Leap has added support for DR services for workloads running on Nutanix private clouds using VMware ESXi.
Nutanix is adding Xi Leap availability zones in Italy, Japan and Germany in Q3, 2019. They join availability zones in US West, US. East and the UK.
Xi Beam is Nutanix’s cloud governance software. The SaaS application will include a real-time security compliance module to identify critical cloud infrastructure vulnerabilities and recommend specific remediation.
Real time security compliance capabilities in Xi Beam are available to customers in early access, and general availability is expected in Q2, 2019. Xi Frame for Nutanix Enterprise Cloud Platform deployments using AHV is available now.
Fujitsu Nutanix HCI
Fujitsu has announced the global rollout of a combined hyperconverged system with Nutanix.
The combo is already on sale in EMEIA and is based on Fujitsu’s PRIMERGY servers running Nutanix HCI Enterprise Cloud software. The system supports VMware ESXi and the free Nutanix AHV hypervisor and scales from a single node, with no maximum limit.
The system is designed for enterprise data centres, hyperconverged business applications, virtual desktop infrastructures, dedicated block storage and ROBO (Remote Office / Branch Office) scenarios.
It’s shipped ready-to-run and pre-configured, with integrated automatic load balancing and built-in resiliency. Fujitsu Infrastructure Manager (ISM) software provides converged infrastructure lifecycle management from a single console.
Fujitsu is the single point of contact for hardware and software support at infrastructure level, with additional services available, including system optimisation.
Nutanix Enterprise Cloud on PRIMERGY is available to order in EMEIA and North America and will be rolled out globally. It is available directly from Fujitsu and through Fujitsu SELECT channel partners. Pricing varies according to configuration and region.
Lastly, Nutanix is making a full-fledged IOT stack in the cloud available through the Xi IOT Cloud instance.
