Home Tags Posts tagged with "HPE Storage"
Tag:

HPE Storage

DatacenterStorage

HPE Primera Eye-Candy Post

by Philip Sellers
written by Philip Sellers

Discussed the changes between HPE Primera versus the 3PAR 8000 and 20000 series, the architectural differences and software differences.

  • HPE Primera 630 / Dual Controller, single processor, 128GB of RAM, single ASIC per node
  • HPE Primera 650 / Quad Controller array, 4 ASICs per node, single processor per node, 256GB of RAM / HPE Primera 670 array is same form factor with two processors and twice the RAM, either 512GB or 1TB of RAM per node
  • HPE Primera arrays do support SAS expansion shelves to extend capacity / Just a few shelves will max out the capabilities of the array / any more would be processor bound
  • The 4 node HPE Primera 650 and 670 chassis / 48 drives in the 4U chassis
  • Controller nodes from an HPE Primera 670 array / dual Intel processors, RAM, Quad ASICs and 3 PCI cards
  • Interconnects for the controllers go into a mid-plane for the array – similar to a blade chassis
  • 4 ASICs per node on the Quad Controller arrays
  • Three PCI slots per slot, SAS expansion slots below

Disclaimer: HPE paid for my trip to attend HPE Discover. All thoughts and writing in this post are my own views and opinions and not reviewed or dictated by HPE.

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

Looking inside the HPE Primera array’s internals

by Philip Sellers
written by Philip Sellers

With the introduction of HPE Primera, lots of questions come up about the internal components and makeup of the new storage array. Is it Nimble or is it 3PAR? Is it all-flash or is it NVMe? Can it support spinning disks? This post is a living doc (I’ll update as I get questions or answers).

Is it Nimble or 3PAR?
That’s a tricky question – however it shares a lot with 3PAR but marries the extreme capabilities of 3PAR arrays with the simplicity of Nimble and the deep integration of Infosight. While 3PAR has Infosight support, it is not complete like Nimble’s integration.

Is it all-flash? Is it NVMe?
According to staff, it is flash storage on SAS in the array. It is not NVMe flash storage. NVMe uses the PCI bus to take advantage of the native capabilities of flash storage, but the NVMe transport lacks a lot of maturity and capabilities needed in enterprise storage arrays – like multipathing – so this is not a shocking revelation. NVMe will come in the future and the arrays are engineered to support it.

Does it support spinning disks?
Why yes, it does. But it does not do automatic tiering like the 3PAR arrays before it. Spinning disk can be added to the array as a discrete tier and could be used for archival. With a strong focus on Tier 0/Mission Critical targets, its unsurprising that the company chose to remove some data services, like tiering.

Is the array self-installable or does it require services for install?
HPE Primera arrays are self-installable. In line with the goals to make the arrays radically simpler than any other mission critical storage array, the 4U chassis is able to be installed by a customer and no longer requires services.

What is the Timeless Storage feature of HPE Primera?
Seems like HPE is chasing Pure Storage and their Evergreen refresh model with HPE Primera. Timeless Storage allows customers to do in the field, non-disruptive controller upgrades. No more fork-lift upgrades according to the marketing.

What is the Services-Centric OS feature of HPE Primera?
HPE Primera’s OS has been architected to decouple the data services from the core OS – much like microservices architecture in modern, cloud-native applications. This allows HPE to release and upgrade software on a continual basis, without disruption.

Does Primera miss any features that are in 3PAR?
HPE Primera does not support native File capabilities. The first release of Primera will not have VVOL support.

Disclaimer: HPE paid for my trip to attend HPE Discover. All thoughts and writing in this post are my own views and opinions and not reviewed or dictated by HPE.

1 comment
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

HPE Primera is company’s new mission critical data array built for simplicity

by Philip Sellers
written by Philip Sellers

This week at HPE Discover in Las Vegas, HPE announced a new storage platform, HPE Primera, built on the technologies of both its Nimble and 3PAR storage lines. The new platform is targeted for mission critical workloads and offers a 100% uptime guarantee for customers.

The new array draws a lot of from the 3PAR line of storage arrays, with scale and speed in the new all-flash Primera. There is shared heritage and data services between the two platforms and HPE is committed to 3PAR in the mid-market for the future – meaning these two lines will co-exist for some amount of time. The portability of data using Peer Copy as well as the ability to protect data with Peer Persistence (aka Metro Storage Clustering for VMware folks) mean that these arrays can protect your data to provide the 100% uptime guaranteed by the company.

In terms of the Nimble heritage for the array, two key features made their way into Primera – Infosight and Simplicity. Gone are the legacy 3PAR constructs of CPG creation and other base setup — the Nimble simplicity is front and center in Primera. While the introduction of 3PAR SSMC (StoreServ Management Console) went a long way to simplifying the complexity under the covers of 3PAR, Primera is said to go even further in simplification towards the startup and administration model of Nimble.

Infosight plays a primary role in the Primera arrays. While HPE has worked to bring 3PAR and Infosight together, the company has native support in the Primera platform – taking advantage of building a brand-new array and platform. In fact, Infosight is large component of how HPE is delivering the 100% uptime guarantee.

In terms of guarantees, the 100% uptime is a bit disappointing. While it sounds great in terms of guaranteed uptime, the penalty of a miss is not substantial for the company. The 100% uptime offers customers up to 20% of the cost of the array back for use with upgrades and future purchases. So essentially, the company if offering store credit for your money spent. I get that it is value, but its not a pure guarantee like most other SLA guarantees.

HPE says the Primera storage arrays will be available for ordering in August, 2019, but no pricing was announced. For more information, see https://hpe.com/storage/Primera.

Disclaimer: HPE paid for my trip to attend HPE Discover. All thoughts and writing in this post are my own views and opinions and not reviewed or dictated by HPE.

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

HPE introduces Optane in 3PAR and Nimble, skips over NVMe drives

by Philip Sellers
written by Philip Sellers

This week I’m onsite at HPE Discover and I have had some interesting discussions with folks from HPE about NVMe and their choices around array architecture in 3PAR and Nimble.  HPE announced Memory-Driven Flash that is available for 3PAR arrays now and coming to Nimble arrays in 2019.  Memory-Driven Flash is storage class memory as a caching layer in its existing array architectures and is based on the Intel Optane storage class memory in an add-in card (AIC).  3PAR and Nimble are not utilizing NVDIMM forms of Intel Optane in the release, though that difference should not be a detriment in the current designs.

For the management at HPE, their talking points are around latency analysis and pinpointing the bottlenecks within the IO pipeline.  Both product management teams (3PAR and Nimble) insist that their current bottlenecks are not around the backend storage, but rather the front-end controllers in today’s arrays.  So simply enabling NVMe based drives in the array does little to address customers real concern, which is lower latency storage IO.  With controller bottlenecks in play, a lot of this  ceiling is created because of the extra processing of data services on the controllers.

In testing, the teams discovered that using storage class memory as a caching layer – which the company showed as a technical preview in 3PAR last year at Discover – greatly accelerated workloads and reduced latency.  The released products are based around the same design as the technical preview.

So why not also enable NVMe drives at this point?  HPE says it is waiting out the emergence of standards and protocol maturity around NMVe before moving to NVMe drives – and because the backend is not where they see the latency introduced in the storage stack.  They reiterate the point that the backend drives are NOT where arrays are constrained – and that is an industry-wide assertion, as I understood it – but rather at the controllers.

NVMe as a protocol lacks support for multipathing and failover and other essential resiliency features taken for granted in the SCSI protocol.  The short answer is simply that SCSI is robust and NVMe is new with lots of growth.  From personal testing of an NVMe over Fabric array, driver support is spotty at best with just a few Linux distributions offering support for NVMe over Fabric.  Microsoft has taken the stance that it is not even offering driver support at this time for NVMe over Fabric.  The lack of multipathing means that NVMe over Fabric LUNs cannot be used for things like VMware VMFS datastores or clustering use cases – many of the high-profile, high-importance use cases typically associated with super-fast, low latency storage arrays.

That led me to ask them about their data services adding to latency that might otherwise be avoided in future iterations of their arrays.  Product managers agreed and said that was the thinking behind the introduction of adaptive data services in the Nimble platform.  By making the data services selective, you can achieve the full potential of the hardware, but you also have the choice to enable those data services where desired to improve the cost-effectiveness of the array.

1 comment
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

Do you own HPE 3PAR and StoreOnce? You should look at Recovery Manager Central

by Philip Sellers
written by Philip Sellers

For every StoreOnce customer out there who also uses 3PAR storage, I think you owe it to yourself to take a look at HPE’s Recovery Manager Central (RMC) product and specifically at its Express Protect.  RMC Express Protect offers the potential of significantly faster backups, longer term retention, and a huge improvement to the workflow and capabilities of your backup and DR strategies.

What is RMC?

Recovery Manager Central is a product that has existed with 3PAR for a long time.  It provides the ability to take application consistent snapshots with a 3PAR array – utilizing built-in software features, like VSS on Microsoft, to coordinate the snapshot ensuring data integrity.  HPE currently offers RMC for VMware vSphere, Oracle database, SAP HANA, Microsoft SQL and now Exchange.

Express Protect is a feature of RMC that offers direct integration between the 3PAR and StoreOnce arrays allowing 3PAR snapshots to be offloaded directly to StoreOnce backup storage arrays for longer term backups than array-based snapshots allow using deduplication and Catalyst compression.  Once a volume is protected with Express Protect, only the delta changes of a volume need to be transmitted to StoreOnce to create a new recovery point.

HPE says that they are seeing up to 23x better backup speeds and up to 15x better restore speeds compared to traditional backups.

Why is RMC so good?

RMC is just an enabling product, to me, in a customer environment that already includes StoreOnce and 3PAR.  It provides the application consistent snapshots needed for data integrity in a DR or backup scenario on 3PAR arrays.  And RMC unlocks a native pathway between 3PAR and StoreOnce for direct, fast backups.

With the current threat landscape from ransomware, it can be hard to determine when an infection occurred.  To lessen the potential impact of a ransomware attack, a lot of companies are turning to backups and snapshots as mitigating steps.  Express Protect gives you both capabilities in the same package – on steroids.  Its faster both to backup and restore than traditional methods.

Native replication exists in both the 3PAR and StoreOnce platforms, meaning you can easily leverage this technology in multiple locations for DR purposes.  StoreOnce provides a number of other alternatives that 3PAR doesn’t have.  Because StoreOnce is software defined, customers can spin up a low-cost StoreOnce virtual appliance in another datacenter on cheap-and-deep storage or even in Microsoft Azure.  This can become an offsite replication target.

In DR scenarios, your replicated data can be mounted as a read-only copy from the StoreOnce and data can be pulled and restored from this read-only copy – but this will be SLOW.  StoreOnce is not intended to run primary storage workloads and the read-only copy would allow for a restore but not for a workload to run.    This means you wouldn’t need a 3PAR on the remote end, however it would work best with a 3PAR on the remote side.

HPE also has an offering called Cloud Bank Storage, which is cloud object storage as another target for StoreOnce replication.  This can be Amazon Web Services object store, Azure object store or an on-premises Scality object store as a target.

Integrations with third-party software vendors, like Veaam and Symantec, can also leverage native capabilities of StoreOnce and RMC to orchestrate jobs.

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
CloudDatacenterStorage

3PAR gets InfoSight and HPE InfoSight gets an upgrade

by Philip Sellers
written by Philip Sellers

HPE has announced support for 3PAR storage with its HPE InfoSight cloud analytics product, acquired with Nimble, along with new updates to InfoSight focused on proactive recommendations.  Even before the conference officially begins, HPE is releasing announcements that customers can dig into more during the Discover event in Madrid this week.

As promised earlier, the InfoSight support for 3PAR is  coming to all existing customers with active support contracts on their 3PAR arrays.  Since 3PAR already had a rich set of telemetry built into the platform, the main hurdle of integration was letting the InfoSight Data Scientists loose on the raw data to map and make sense of it.  HPE leveraged existing code and analytical models against the 3PAR data to do this quickly [HPE completed the Nimble acquisition just 7 months ago].

The first iteration of InfoSight for 3PAR includes three feature sets:

  • analytics between the storage and virtualization platform to find root cause of problems that existing between storage and virtual machines
  • a unified cloud portal combining the HPE StoreFront Remote capabilities with InfoSight for trending, capacity predictions and health checks
  • the foundations of predictive support, which in the future will eliminate most tier 1 and 2 support cases, instead routing cases directly to level 3 (aka. adoption of the Nimble Storage support model for 3PAR)

Not resting on past accomplishments, the InfoSight team also has a new set of capabilities it is unleashing to customers – Preemptive Recommendations.  For these, HPE is drawing insights from the data it is getting from the storage and virtualization platform to make recommendations that prevent issues, improve performance and optimize the environment.

To a large degree, what HPE is talking about in preemptive analytics in InfoSight sounds a lot like the problems VMware is trying to solve in vSphere with proactive recommendations from vRealize Operations Manager (vROPS).  Where these recommendations can potentially shine in InfoSight are in the deeper insights that correlate between the hardware and the software since InfoSight has deeper visibility into that platform.  Clearly, all tech companies are realizing that there is tremendous value in machine learning applied to IT support.

HPE is using machine learning and expert defined conditions to generate preemptive recommendations.  Since the learning engine is cloud based, the refinement and self-learning should power a number of wins for customers as the install-base widens and customers adopt this support tool, making it difficult to replicate for on-premises prediction engines.

0 comments
1 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

Post acquisition, Nimble is finding its place in HPE’s portfolio

by Philip Sellers
written by Philip Sellers

I had the chance this week to attend HPE’s Storage Tech Day in Anaheim, CA.  I’ve been fortunate to get updates about the storage portfolio and direction during HPE Discover events each year, but its been a while since I’ve had a full day devoted just to storage topics with HPE.  Going into the event, I hoped that it would clarify the position of Nimble in the storage portfolio and I was glad to come away with some answers.

HPE is quick to point out that in acquiring Nimble, they acquired Infosight, the cloud analytics platform that Nimble had created.  Infosight is not just an analytics tool for the Nimble storage arrays.  Nimble realized early on that many of the perceived “storage problems” were never actually the storage array.  As a startup trying to make its way in the market, it needed a way to help diagnose and pinpoint these other problems.  Thus, Infosight was born out of the Nimble support organization.

Infosight combines the telemetry it gets from Nimble storage arrays along side of data it extracts from vCenter using vSphere API’s.  It correlates this data and tries to identify patterns of activity not only for a single customer, but for its entire customer base.  The Nimble team gave one great example of the power of these correlations.  A customer called in with a problem that was determined to be an iSCSI issue in vSphere 6.5 that occurs under certain circumstances and with certain vSphere drivers.  With this data, Nimble was able to identify all of the customers who fit the same conditions and issue a proactive alert to them and also work with VMware directly to address the iSCSI bug that was found.  Incidentally, although I’m not a Nimble customer – I was one of those customers affected that Nimble helped VMware to fix.

Nimble quotes impressive numbers when it comes to Infosight.  Some 86% of all cases for Nimble are autogenerated and solved by Infosight.  And 54% of cases are not actual Nimble array issues at all.  In fact, Nimble says Infosight is such a good analytic tool that many of its customers use it as first line support for more than just storage problems.  (Is that a good thing?  Yes, it proves Infosight is good, but… )

So, point #1 – HPE acquired Infosight and has big plans for it – so much so that its already branded HPE Infosight.  The first order of business is getting Infosight and 3PAR married together.  HPE says this is coming in the near future, but stopped short of quoting timelines.  3PAR already has rich telemetry within its arrays and has since its inception, but there is still work to be done to integrate it.  HPE’s StoreFront Remote will be folded into Infosight at some point, due to overlapping capabilities.  HPE also warned that we shouldn’t expect full capabilities in the initial release of 3PAR support for Infosight.

But, where do Nimble arrays fit?

But beyond the analytic cloud that Nimble created, Nimble also makes storage arrays.  They seem to be good arrays (never have run them, so I say seem), but there isn’t a striking differentiation between the Nimble Hybrid Flash or All Flash storage arrays and any other dual controller array on the market – except for Infosight.  So that begs the question, where does it fit in HPE’s portfolio.

For that answer, there are a few things I took away from the event this week.

Nimble All-Flash is a very competitive flash offering in the small to medium business and enterprise markets.  It is priced below what HPE can sell a 3PAR array.  It makes it an excellent entry all-flash array, also.  That’s not to pigeon hole the array, because it fits nicely in many larger enterprise business, also.   Nimble All Flash provides deduplication and compression, as expected with any all-flash array.

Nimble Hybrid-Flash is highly competitive in the small to medium business markets.  These make great lower tiers of storage in enterprise customers and fantastic starting points for small customers with less demanding IO requirements.  Its much better than an MSA, but below a 3PAR.  HPE admits there is overlap, but they don’t see it as a bad thing – they simply see it as choice.

And then, there is the Nimble Secondary Storage arrays.   Targeted at backup, I knew the least about these going into the tech day.  A Nimble Secondary Storage array is simply a Hybrid Flash array with global dedupe.  The downside to enabling global dedupe is performance impact – the Secondary Storage array is slower than the Hybrid-Flash array, however, this makes it well suited to backup applications and deduplicating mostly-cold datasets.  It also provides higher performance than most disk-to-disk storage arrays, like StoreOnce, allowing flash-boot capabilities and the ability to run some workloads on the array for recovery or dev/test – directly from the backup repository.  This is a killer feature for many virtual backup solutions like Veaam, Nakivo, Vembu and others.

Disclaimer: HPE provided the travel, lodging and meal expenses for my trip.  The information in this blog post comes only from me based on the observations and information shared during the Tech Day event, but was not dictated or reviewed by HPE.  

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

New 3PAR 9450 offers higher performance & scale for the mid-range

by Philip Sellers
written by Philip Sellers

A few weeks ago, HPE introduced a brand new 9450 all-flash model of its 3PAR storage arrays leading up to HPE Discover and the company showed off the new arrays during the event.  The 9450 joins the current offering of 8000 and 20000 series arrays and is the first in a new 9000 series of arrays.  The new model is targeted at the mid-range, the same as the 8000 series arrays, with up to 4 controllers in an active mesh configuration.   HPE is quick to say that the 9450 is not a replacement for the 8450 all-flash model, positioning the new array for customers who want larger scale, the ability to utilize NVMe and higher performance.

The new 9450 is an all-flash offering and HPE is saying that it will deliver double the scale and double the horsepower of its 8450 with the 9450 array.   The 9450 is based on the same HPE Gen5 ASICs that power the 8000 series arrays but will offer greater scale for the processors that power the array and more DRAM for operation.  In terms of resiliency, HPE says the new array boasts six 9’s availability.

Digging into the announcement, what are the highlights?

  • Scale to 18 PB of usable all-flash capacity on 6PB raw capacity (standard 3 to 1 efficiencies)
  • 1.8 million IOPs at sub millisecond latency (70% increase)
  • Up to 34 Gb/s in bandwidth
  • Storage Class Memory and NVMe ready
  • 80 host ports, more than 3x more than the 8450

Details on the processors is not yet available, so an assumption can be made that it will be the new Intel processors set to release in the coming weeks.

Density Differences

 

 

The 3PAR StoreServ 9450 utilizes the same controller design as the 20000 series arrays.  This 8U chassis for its 4 controllers lacks any SSD drive bays, meaning that the 8000 series models with controllers integrated into a drive shelf can accommodate more SSD drives in less rack units.  The trade off is the ability to accommodate NVMe storage in the controller on the 9450 which the 8000 series controllers do no have space to accommodate, meaning in the 9450, you can get some storage in the controllers.

With the new 9000 series and 20000 series controllers, they’re a 2U size, meaning more room for processors, memory and half-height PCI cards.  All of this equates to the ability to push the high-end limits of the mid-range 3PAR higher than ever.

 

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

HPE is talking about Composable Data Fabric

by Philip Sellers
written by Philip Sellers

Last week at HPE Discover, HPE introduced the idea of a Composable Data Fabric on the show floor to customers. While there was no specific product announcement around the concept, the idea is HPE’s vision for the future of software defined storage.  The overall strategy will utilize HPE OneView for management and control along with HPE StoreVirtual as the storage platform.

Composable Data Fabric is the idea of federation and data mobility within software defined storage.  Regardless of how you start with StoreVirtual, HPE is working to create the ability to move your data from one StoreVirtual system to another.  This will offer its customers the ability to migrate their data seamlessly and nondisruptively to and from any form factor to another.  The solution will be hardware agnostic since it is based on HPE StoreVirtual – meaning it can be installed on standard HPE ProLiant servers, purpose built HPE StoreVirtual hardware, StoreVirtual VSA on ESXi and inside of Synergy Platform, HyperConvered and ConvergedSystem solutions.  One of the main goals is to ease the disruptive and difficult nature of data migrations.

Composable Data Fabric is also about enabling all the needed services within StoreVirtual – specifically adding the ability to do file services and object storage, in addition to block.  In the near future, HPE is saying that StoreVirtual will be able to natively host file services, the same way that 3PAR StoreServ arrays are able to.

The Composable Data Fabric idea continues on the path of growth and innovation to the StoreVirtual line of storage.  Within the Synergy Platform composable infrastructure strategy, StoreVirtual plays the key role to unlocking local disks within one frame and exposing that data storage to systems throughout the rest of the data fabric or network connected to the Synergy.   Because StoreVirtual allows Synergy to become its own storage array, Synergy can become a self-contained platform for hosting any application – whether its cloud-native, virtualized or traditional client-server, with the ability to flex and change as requirements change.

The primary takeaway from the announcement is that customers will enjoy complete flexibility and mobility for their data between StoreVirtual installations – regardless of the shape or size.  By mixing different instances of StoreVirtual with different performance capabilities, its easy to envision a great software-defined storage with tiers of storage based on performance or cost.  Add the ability to move data seamlessly between them as the requirements of the data change, and that becomes a winning combination.

 

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
DatacenterStorage

HPE unleashes monster sized SSD drives for its 3PAR StoreServ arrays

by Philip Sellers
written by Philip Sellers

HPE announced massive new SSD drive offerings for its 3PAR StoreServ arrays — a 7.68TB and 15.36TB drive in size. The new drives double and quadruple the former 3.84TB high-end drive. And with these new drive offerings, HPE has leapfrogged the spinning disk capacities with this new generation of SSD drives with incomparable performance.

Sizing

The first thing you may notice is the odd sizing on these drives, but those sizes are the product of taking industry standard SSD drives  and applying the 3PAR method of sparing to them.  The previous large capacity SSD for a 3PAR StoreServ was 3.84TB in size, also an odd size.  Truth be told, these are the same enterprise drives that the SSD vendors sell as a 3.2TB drive.  The 3PAR arrays don’t actually use a special type of SSD drive – these are the same standard SSD drives available to other vendors.  So how do you get this odd increase in size?

With the 3PAR StoreServ architecture, it has always allocated spare space at the array level, reserving space for failures.  This goes back to the origins of the 3PAR’s design on spinning disks.  Because 3PAR has a sparing technique called Adaptive Sparing at the array level, HPE went to the SSD vendors and asked them to allow HPE the reserved space on the SSD normally used as spare blocks.  By combining Adaptive Sparing across a much larger set of disks, HPE is able to unlock access to to more of the capacity you’re already paying for and let you get real use out of that.  HPE is able to do this without introducing new risk because it is already handling sparing at the array level.  Where HPE could have been at a disadvantage due to it’s sparing technique [taking away space from a 3.2TB, 6.4TB or 12.8TB disk], this becomes a benefit to the customer.   HPE also stands behind its SSD technology with a 5 year, unconditional warranty on its SSD drives.

Now, this year, another thing has happened.  Since HPE went to the drive vendors and asked to tap into the spare space, the drive vendors themselves are sizing the enterprise SSD’s at the newly unleashed larger sizes and other storage vendors are following the HPE lead in handling sparing at the array level.  So, in this case, you’ll see Samsung advertising the 7.68TB and 15.36TB drive sizes – instead of 6.4TB and 12.8TB drives with full fixed overprovisioning space. I talked to a Samsung marketing manager while at HPE Discover and he confirmed these sizes.  (That’s a whole lot of spare space as we keep sizing up, isn’t it!?)

The Real Story Isn’t the Sizes

Folks I talked to at HPE Discover in Las Vegas last week insist that the real story isn’t the increased drive sizes.  The real story is around a new technique in Adaptive Sparing – something HPE is calling Adaptive Sparing 2.0.  To really understand the benefits of the new technique, its important to know that SSD’s work on a ‘erase before write’ principal, explained in the graphic below.  Adaptive Sparing 2.0 opens up more free space to accelerate writes on each drive.

WhatsNew_Page_12

WhatsNew_Page_14

Adaptive Sparing 2.0

The Adaptive Sparing 2.0 technique combines free space, spare space and the fixed overprovisioning space on the SSD together to further increase the effective overprovisioning space of the array.  Fixed overprovisioning space is the fancy term for that spare space the SSD vendors reserve on SSD drives.

 

3PAR StoreServ arrays also combine other features like system-wide striping and express layout to allow for greater performance and endurance of the drives.  These features allow for even use across your SSD drives which prolongs the life of the drive wearing the same few drives out prematurely.  These same features that boosted performance by spreading data across a larger number of spindles on spinning disk is also benefiting you in the SSD world today.

The full active/active nature of the 3PAR architecture also allows increased performance – with active/active host connections for optimal performance host to array, mesh-active controllers where all controllers can access all the drives in the array for even wear and balance of data, and for SSD drives, each drive is dual-ported, allowing active/active SSD access from each node in the node pair.  Dual-ported SSD drives doubles the available bandwidth to and from each SSD – allowing for full duplex speeds for every IO transaction.

Whether by incredible luck or very strategic planning or most likely a little of both – the truth is customers are still benefiting from the architecture choices made well over a decade ago when the 3PAR arrays were first being designed.

HPE has a slide deck with more details that I’ve embedded below:

 

0 comments
0 FacebookTwitterPinterestLinkedinRedditEmail
  • 1
  • 2
@2019 - All Right Reserved. Designed and Developed by PenciDesign