VPLEX / VE Is Here

Funny, one of the most anticipated products (IMHO) just pop out, you can bet you will hear more and more about it in the next months to come, VPLEX / VE was just a part of one of the biggest (if not THE biggest) DPDA launcher ever..

the theme was Data Protection for a Software-Defined World.

image

There are 3 specific sub-themes that support the overall launch. These sub-themes are key requirements for delivering effective data protection in a software defined world.

First – Delivering “data protection as a service” supports the shifts that our customers are making toward offering “IT as a service” – Effectively, becoming a service provider to the business. This also extends the opportunity to EMC Business Partners to offer “data protection as a service” to their end user customers.

Second – Empowering data owners – and by this we mean application, virtual and storage admins – with visibility and control of their own data protection needs, from familiar interfaces, removes the need for data protection silos and attacks the problem of “accidental architectures”.

And finally – Seamlessly spanning “The Continuum” of data protection, to assure appropriate service levels, based on the value of the data within the applications, from continuous availability to replication to backup and archive

Most industry analysts agree that there’s a need for a tiered recovery plan, in order to meet multiple services levels – from zero downtime and no data loss, to various point in time copies, to secure long term retention with archiving.

image

VPLEX 5.4 and VPLEX Virtual Edition are important releases for several reasons:

• VPLEX/VE is a new entry point for application data availability and mobility. Reduced TCO derived from deployment on standard ESXi server hardware and low cost iSCSI infrastructure. VPLEX/VE is a VMware-centric solution with deep VMware integration, managed and monitored through vCenter.

• VPLEX 5.4 adds support for MetroPoint Topology – Providing multi-source replication to a 3rd site – extending the VPLEX and RecoverPoint continuous availability story to include protection against events that could impact two sites, and operational recovery from data loss or corruption with any point in time RecoverPoint functionality.

• In the same timeframe as these VPLEX releases, ESA 2.3 for VPLEX delivers new predictive analytics capabilities, for VPLEX and EMC storage attached to VPLEX.

image

No competitive solution offers the continuous availability and protection that MetroPoint delivers.

Users get what they want – a single DR copy of their data, multi-site protection, no impact even though failures of two sites.

Conversation opportunities –

•New incremental revenue from installed VPLEX customers as they improve their data protection to both sides of a VPLEX Metro topology

•For new customers, you have the most comprehensive continuous availability and data protection story to tell

image

VPLEX/VE is a software based, simple, affordable, virtual storage availability platform that provides continuous availability and non-disruptive mobility targeted at mission and business critical applications in the midrange and enterprise organizations with VMware and Mid-tier iSCSI storage infrastructures

image

VPLEX/VE takes the same great technology and the associated use cases the industry has come to know and trust and delivers them as a set of virtual appliances in a fast and easy to install VMware vApp package.

This is unique in the industry as it is the first and only continuous availability platform that is designed to run natively on the customers VMware infrastructure.

This simple, affordable and virtual storage availability platform is specifically designed midmarket customers who are looking to achieve the highest levels of availability for applications running in their VMware environments running on iSCSI storage infrastructure.

Now the storage availability conversation can extend beyond the storage admin to include the vAdmin who will directly benefit from improved availability at the VM level.

image

VPLEX/VE is integrated into vSphere

•Manage through a single tool

•Leveraging existing VMware expertise

•Provision as part of VM creation

•Use with your automated work flows

•One stop shop for all event info

•Eliminate multi-GUI for monitoring and troubleshooting

Benefits of VPLEX/VE and vSphere integration

•Deliver on infrastructure demands faster

•Faster and easier provisioning as part of VM creation process

•Shortened learning curve for admins who know vSphere

•Improve incident resolution

•Faster problem resolution with fewer, more powerful troubleshooting tools

Provision and Manage Storage With VM Workflows

•Provision highly available storage resources as part of the VM creation process

image

VPLEX/VE is integrated into vSphere

•Manage through a single tool

•Leveraging existing VMware expertise

•Provision as part of VM creation

•Use with your automated work flows

•One stop shop for all event info

•Eliminate multi-GUI for monitoring and troubleshooting

Benefits of VPLEX/VE and vSphere integration

•Deliver on infrastructure demands faster

•Faster and easier provisioning as part of VM creation process

•Shortened learning curve for admins who know vSphere

•Improve incident resolution

•Faster problem resolution with fewer, more powerful troubleshooting tools

Provision and Manage Storage With VM Workflows

•Provision highly available storage resources as part of the VM creation process

image

VPLEX/VE provides benefits to a different set of customers than the traditional appliance based VPLEX. VPLEX/VE will be particularly valuable to the following customers.

Customers with VMware environments and iSCSI networks. VPLEX/VE supports iSCSI networks but not FC networks. iSCSI networks are common at SMB customers and branch offices of enterprises because of the lower implementation and maintenance costs for iSCSI. VPLEX/VE is tightly integrated with Vmware and is managed through the vCenter GUI. All management activities such as provisioning, moving volumes and monitoring storage is done through the VPLEX/VE plug in for vCenter. Customers no longer need to develop new skills in separate management tools.

VPLEX/VE benefits customers who want to instantly move VMs between sites. Unlike Storage vMotion, VPLEX/VE continuously mirrors application data (and VM data) between sites so that when a VM Vmotion is directed by an administrator, the VM can move to the new site without delaying for VM data to be copied to the new site. And because this is an active/active mirror of data, VMs can be moved back and fourth dynamically. Storage vMotion requires VM data to be copied to the new location first, before the VM can be moved often taking hours to complete a Vmotion operation. And each Storage Vmotion is a new activity so if a VM needs to be moved back, the VM data needs to be copied back.

Similarly, customers who want to load balance between sites can do so instantly and dynamically because the application data and VM data is mirrored between sites.

All of this is enabled by VPLEX/VE’s ability to mirror application data and VM data between arrays at a single site or between two sites. In addition to the added mobility benefits, customers who want to maintain mirrored data between arrays at one or two sites for improved availability can use VPLEX/VE to accomplish this.

image

VPLEX/VE leverages many of the unique capabilities of VMware such as Vmware High Availability (HA) to help create a continuously available application environment.

VMware HA provides the ability to automatically restart VMs when a physical ESX host hosting those VM’s suffers a failure. Under these circumstances, VMware directs a new ESX host to reboot the failed VM automatically using the same VMDK data stored on the single array.

The challenge is that VMware HA requires a single source of VMDK data. So in the loss of an array or in the loss of a full site, VMware HA is unable to restart servers.

image

VPLEX/VE bridges the gap in VMware HA by mirroring VMDK data across two arrays in two sites but presenting that data as a single volume to ESX hosts.

In the even of an outage from something like a site failure, VPLEX continues to make that VM data available at the remaining site.

And the ESX cluster is able to AUTOMATICALLY reboot the VMs on the remaining site without any human intervention.

image

While continuous availability is the key use case for VPLEX/VE and VMware, VPLEX/VE also enhances data mobility for VMware environments beyond what can be achieved by Storage vMotion and Storage DRS.

Storage vMotion gives vAdmins a way to non-disruptively move their VMs from one array to another. This can be done in a single site or across sites.

But Storage vMotion will not complete the move of a VM until the VM data move is actually completed. As a result, VM moves can be delayed. This is particularly challenging when multiple VMs are being moved simultaneously.

In addition, when a Storave vMotion move is requested, the ESX Host resources can quickly be overloaded by the sudden burst of CPU and resource activity as the VM data is replicated by the ESX host from the old array to the new array.

image

VPLEX/VE operates by continuously mirroring data across arrays so that the exact same data is read and write accessible from two different arrays simultaneously. When a vMotion request is made that requires the VM data be located on a new array such as when VMs are vMotioned across distance,

The vMotion move can be completed instantly because the mirror of the VM data can begin to be used as soon as the VM is moved to the new location.

image

Distributed Resource Scheduling can load balance VMs across your ESX cluster for best use of your compute assets. When DRS requires VM data to be moved to a new location to complete a DRS load balancing move, it uses Storage DRS, a process very similar to Storage vMotion.

And just like Storage vMotion, Storage DRS does not actually move the VM until the VM data is replicated to the new array.

Also like Storage vMotion, the process of moving data with Storage DRS can consume a lot of ESX host resources in a burst fashion to conduct the load balancing VM data moves.

image

Again, because VPLEX/VE mirrors data and makes it read and write accessible from multiple locations simultaneously,

VM moves for load balancing with DRS happen instantly and do not spike the ESX host resource usage in the process.

 

VPLEX Virtual Edition Cluster Witness Demo

Leave a comment

Filed under Uncategorized

EMC World 2014 Is Coming ! Here’s the XtremIO Sessions

image

Hi,

EMC World 2014 is coming soon, you can click on the banner above to register for it, this year is going to be VERY interesting as any other year but from our perspective (the emerging technology unit), it will be DIFFERENT, no more “area 51” about this “future” AFA that WILL change the world, this year, the XtremIO array has already been GA’d and it’s already CHANGING that market, many production customers are already USING the product and are amazed by the value it is bringing them every day and as such, we will have a lot to share, attached below are the sessions we have planned, you are all welcome to register for them, yours truly will have the VDI session running twice (or so I have been told) and of course, we would love to interact with you, our existing and future customers!

image

image

image

image

oh, and EMC is notorious for product announcements..

Smile

Leave a comment

Filed under Uncategorized

XtremIO 2.2 Service Pack 3 Is Out!

Hi,

If you thought 2013 was big for EMC and XtremIO, you have seen nothing yet,

2014 is going to be much much bigger, a big part of the launching theme was around to “getting the architecture right!”, this was well needed because as we said, features are easy to add, changing your core architecture isn’t!

as part of let’s add some features, we have just released the 2.2 Service Pack 3 upgrade for our XtremIO customers,

this is what i call, “A minor release” hint, hint, the next one is kinda big..

Here’s whats new:

Hardware Changes:

clip_image001

20TB Brick

Enables XtremIO to provide more capacity in a single cluster for non-dedup use cases like Databases. Can be used for VDI as well, when physical capacity is more important then logical capacity.

Based on 800GB Hitachi Sunset Cove “Type B” (Encryption ready)

Supports 1, 2 and 4 X-Bricks per clusters.

Security

Active Directory Integration

clip_image003

Supports AD integration over LDAP and LDAPS protocols

• Allows mapping of AD groups to XMS roles

• Mapping is done via groups, where AD groups are
mapped to roles within the XMS

• Supports multiple AD servers

• Use local store for user authentication if AD service is
not available

Client to XMS security

Allowing only SSH and HTTPS secure protocols.

HTTPS capabilities

– Secures GUI download from XMS

– Encryption of GUI to Management Server communication

– Supports installing 3rd party server certificate

– Allows encryption of the RESTful API

Remote Syslog Support

clip_image005

• Support sending events to a remote syslog server or multiple syslog servers

• Events handler enables the user to set rules on which event should be sent

Management Improvements

clip_image007

clip_image009

clip_image011

Monitor – display 30 minutes of history in line widget

Object Granular Monitoring – add latency reporting

– Average latency per volume

– Average latency per Initiator / Target, per block size

Support VAAI Thin Primitives

clip_image013

clip_image015

1. TP STUN

Return TPSTUN if a write is received by the XtremIO array and it cannot be stored due to lack of physical or logical capacity.

TPSTUN is enabled by default on all volumes

  1. TP Soft limit

A warning is raised and surfaced in VMware vCenter™ via VAAI if a thin-provisioned datastore reaches a specific threshold.

The following CLI commands were changed/added:

modify-cluster-thresholds  vaai-tp-limit=[1-100/NO-LIMIT] 

modify-volume vol-id=vol-id vaai_tp_alerts=[enabled/disabled] 

add-volume <list-of-parameters> added an optional parameter  vaai_tp_alerts=[enabled/disabled]

Leave a comment

Filed under Uncategorized

Virtual Storage Integrator (VSI) 5.6.2 Is Out

 

Hi,

we have just released VSI 5.6.2, i highly recommend this release if you are using it to register VM’s with XenDesktop..

image

as always, download it from support.emc.com

Leave a comment

Filed under Uncategorized

Proud To Be Selected As EMC ELECT 2014

Hi,

I’m so proud and humbled to be selected as an EMC ELECT for 2014, i always thought that sharing information is fun and useful for everyone and i’m happy to get a recognition for it.

Elect2014-web

coincidentally enough, i also received the SE MVP Award for the XtremIO BU, following the same rule as above, share and help other teams is not a one time off task and while its not easy sometimes, it’s very rewarding to know that people can count on you, a typical Virgo sign behavior, i suppose

WP_20140113_22_27_16_Pro

2014 is going to be an amazing year for EMC and the Emerging Technologies Unit!

1 Comment

Filed under Uncategorized

Through the Looking Glass

image

“I don’t understand . . .’ ” said Alice. “It’s dreadfully confusing!”
“That’s the effect of living backwards,” the Queen said kindly: “it always makes one a little giddy at first- ”
“Living backwards!” Alice repeated in great astonishment. “I never heard of such a thing!”
-Lewis Carroll

Hi,

As i’m on way for Emerging Technologies our sales kickoff, i wanted to take a look back into 2013 and reflect what  i have learned.

its also important to note that this is just my own personal view and probably my own “distortion zone” that i lived trough 2013. think of this post as a diary writing.

i joined XtremIO back in the beginning of 2013, the offer to do something completely different was compelling, up until now i was a vSpecialist basically responsible for everything VMware and EMC and suddenly, i have become a dedicated pre-sale to the XtremIO product but apart from just doing pre-sale i was giving (the official) responsibility to be a liaison point between our customers to R&D and our product management team, it is very enjoyable to see a customer request making it’s way to our product faster than ever (an example, will be our upcoming software release to XtremIO..)

so..sitting close to R&D has it’s advantages.

from April to November we launched our Direct Availability program (DA), basically running our product at customers sites and learning from potential bugs that may raise, we also took the opportunity to sharpen our software stability so when the GA release was ready (November 14th), we could release a well and stable product.

i think it will be fair to say that we were surprised by the overwhelming success of XtremIO, i mean, we knew we have a winner product that is every different from anything else that is out there but we also considered IDC and Gartner predicting that All Flash Arrays percentage out of the global storage marker will be fairly small, now, it IS fairly small and I’m very fortunate to work for a company that can sell the right product for the right use case, and Hi, in many cases, a VNX or a VMAX are a better product for the use case but what i wasn’t ready for was a term called “Buzzword Compliance”, basically, a customer wants a specific technology because it’s sexy, fresh and it change everything he know about the storage world. now, technically speaking the all flash array may not be the right product for this specific use case but because the customer wants to establish their data centers with the latest and not aged technology, they chose XtremIO.

so many X-Bricks units have been sold in 2013 and while I’m not allowed to publicly share any specific information, im sure it will be discuss in the next analyst / investment call.

what i also observed in 2013 is that the VDI market is booming, i know it’s a common joke to call every year (since roughly 2010) “the year of the VDI” but man, i was involved is some very large successful wins of XtremIO into the VDI use cases, im literally meeting with a mega large customer the week after next to help them design their supper large scale VDI environment.

i have also seen 2013 as the year of the DB’s of FLASH arrays, it started with a local customer that already runs their core DB’s on a dedicated flash tier on an hybrid array and the latency was kind of ok but their business requirement was to reduce the IO wait time in their core DB’s so immediately an XtremIO POC was conducted and and to our surprise the core DB’s wait time was amazing but the overall transaction layer didnt change, after a further examination, we noticed that the rest of the VM’s in this environment couldn’t keep up and since then, im seeing this in many other sites as well, FLASH changes everything and as such, you should holistically look at your environment and see where would FLASH give you the best benefit, not just your tipping point use case (core DB etc’)

lastly, the VMworld conference opened our eyes and confirmed our belief that when using XtremIO in a test / dev environment where the ability to provision / decommission many vm’s over time is a good sweet spot use case for us, basically, VMware were given 4 XtremIO X-Bricks for their Hands On Labs (HOL’s) and it was amazing to hear their tech ops team do not need to worry about storage bottlenecks during the conference.

The Raise (And Fall) Of Ziggy Stardust

The Rise And Fall Of Ziggy Stardust And The Spiders From Mars

2013 was also an year when was saw at least one AFA market leader that failed it to the finish line, in the very beginning of 2014 we saw another storage startup starting to close down, while i’m not trying to pick up on anyone in specific (hence why i didn’t mention names), it IS important to understand the “WHY”, i think that there is actually more than one reason but generally speaking, it had to do with either technologies that were relevant couple of years ago but unfortunately, the AFA market has moved on, if your are not consistently looking over your should for what advanced technologies are out there, you are very likely to be left behind and this is a very crowded market to begin with.

the other reason is cash burnout, there are many startups out there who prayed to be bought by the big players and since many (not all) of the large storage vendors already made up their mind avout their FLASH strategy, these startup companies found themselves not reaching Round B or C in their fund raising.

so, what’s going to happen in 2014?

more SSD’s drive will be sold which means a constant reduction of the SSD prices, that will mean an ever larger penetration into the comfort zone of the hybrid arrays, new software revision will come out and a larger integration into the EMC eco system will be made, this is in my eyes, the biggest strength of EMC and it always have been, we never sell a point product without the eco system integration, even at GA we already supported PowerPath/VE, VPELX and VPLEX RecoverPoint integration and this is important for our customers because it means that the integration of this disruptive technology into their hybrid based eco system is EASIER.

What Else?

image

Virtual SANS (VSAN)..

you hear a lot about VSAN with EMC products such as ScaleIO (has been GA for a long time now) and VMware own VSAN product (to be released in 2014) called “Vsan”

these are also very disruptive technologies but they are FAR more disruptive than just All FLASH arrays because while with an ALL FLASH Array, you still get a box and the box acts like a..storage box, with VSAN’s, you procure a software and you are using your own servers and you own HDD’s / SSD’s drives to build an hyper converge infrastructure, in my own opinion, these type of technologies will have to keep up with the matures of the “standard” storage arrays but it will come and FASTER than what everyone predict.

image

is there only one product that will win?

NO, different customers have different opinions about technologies and against, it’s important to understand that there isn’t right or wrong here.

one thing is guaranteed, in 2014 and the following years to come,  the storage market will change it’s face in a way and in a pace that it never has.

ok, but how is that related to the “trough the looking glass” ??

well, 2013 was also the year of the marketing clones, many AFA’s looks similar from a sales pitch perspective but there is a world apart between then in terms of their architecture, some competitors out there will try to show their disadvantages as advantages..in 2014, the GAP will become bigger and clearer.

customers knows best

interesting time..

Leave a comment

Filed under Uncategorized

A Tale of Two Architectures — Engineering for the 99.999% versus the 0.001%

Today’s enterprise storage arrays follow one of five different architectural approaches to clustering their controllers for high availability.  We’ve represented them here in order of increasing complexity (from a “how difficult is it to build” perspective), robustness, and performance.

  1. Dual-controller, Active/Passive Model:  This design uses two controllers.  In normal operations, one of the two controllers owns all the I/O processing and data services management (active), while the second controller stands by (passive), ready to take over in the event of failure.  Some variations allow I/O to be received and transmitted on the passive controller’s host ports, but the controller itself does no I/O or data services processing.  Rather it simply passes the I/O over an internal connection to the active controller, which is responsible for all data services (like deduplication, thin provisioning, snapshots, RAID etc.) and all read/write processing to the underlying flash media.  This requires the use of ALUA (Asymmetrical Logical Unit Access) to define the preferred path to the active controller and the non-preferred path to the passive controller.  There is no customer-facing advantage to this model.  The only advantage is to the vendor because this HA model is much simpler to implement and helps achieve a fast time-to-market.  The big disadvantage, especially with flash, is that the active controller quickly becomes the performance bottleneck while the resources of the passive controller are wasted sitting idle.  Active/Passive HA architectures were the original HA mechanism developed in the 1980’s.Screen Shot 2013-12-29 at 8.36.23 AM
  2. Dual-controller, Dual-Active Model:  This design is an improvement on the active/passive model by using both controllers to perform the full range of data activities.  The distinguishing feature is that each controller is the master for a subset of volumes and stands by to take over its partner’s volumes in case of failure.  This is simpler to implement than full active/active HA (where any volume is accessible on any port on any controller and there is no path or performance preference) and is still a popular model in the enterprise storage marketplace.  The drawback with this model is that any given volume is still limited to the resources of a single controller, and in case of failure performance of the array can be cut in half.  Another issue is that the administrator must pay attention to the workload on each controller and manually assign volumes to maintain balance.  Dual-Active architectures became popular around the turn of the century.                                                                                                                 Screen Shot 2013-12-30 at 8.00.07 AM
  3. Dual-controller, Active/Active Model: This is the most advanced form of dual-controller architecture where both controllers are active in data services and I/O processing, there is no assignment of volumes to controllers, and any host can access any volume through any port on any controller without path or performance penalty.                                                                                                                    Screen Shot 2013-12-30 at 8.02.20 AM
  4. Multi-controller, Asymmetric Dual-Active Model:  This model is similar to dual-controller dual-active except that now there can be more than two controllers.  Volumes are still owned by a particular controller and there is still path and performance preference.  The advantage of this architecture is aggregate scale – higher levels of total capacity and performance under a single point of management.  However, any particular volume is still limited to the performance of the controller that owns it.                                                                          Screen Shot 2013-12-29 at 4.29.07 PM
  5. Multi-controller, N-way Symmetric Active Model:  This is the XtremIO’s approach – and it is both the most sophisticated and has the most customer benefits.  In this model, the storage array scales to N controllers (as of Q4’13 N=8 and we will be increasing N in subsequent releases).  During normal operation, all controllers are actively serving I/Os and processing data services and all controllers evenly share the load.  If one of the controllers fails, the remaining controllers continue to serve all the I/Os without loss of service.  During a failure, N-1 controllers remain to service client requests. All purchased hardware is actively working to service client requests, the system scales linearly with more controllers, and degraded performance is proportional to cluster size.  For example losing 1 out of 4 controllers causes a maximum 25% loss of resources while losing 1 out of 8 controllers causes a maximum 12.5% loss of resources.  This is the most advanced and resilient HA mechanism available today and this architectural model is shared with EMC’s flagship VMAX arrays.  Most vendors are not capable of the engineering sophistication it takes to achieve this architecture.Screen Shot 2013-12-29 at 4.30.10 PM

It’s important to note that these various architectural models are truly deep-rooted architectural decisions and it is practically impossible to start lower on the food chain and move up.  In fact, in the history of enterprise storage we can’t think of a single instance of it happening with any vendor.  You can always add features to a storage array – but architectural decisions have a way of sticking around.

While all the above architectures are HA, the practical difference between active/passive and N-way active are huge.  Let’s assume both systems have 99.999% availability (5 nines) so we can compare apples to apples. In the dual-controller, active/passive model, you have a controller (that you paid for) sitting there doing nothing 99.999% of the time.  Your performance is half of what it could be 99.999% of the time too.  On the other hand, the N-way active model gives you the full performance of all the controllers 99.999% of the time. The aggregate performance of the N controllers (even if N=2) is much higher than what you can get from a single active controller in the active/passive model.

Of course, the argument for the active/passive model is that during the rare event of failure in the active controller, the passive controller will become the new active, thus maintaining the same performance as before the failure.  However, the big problem in this argument is that it comes at a huge cost — you are wasting 50% of your performance 99.999% of the time.  You maintain the same (50%) performance in the event of a controller failure, which has a 0.001% chance of happening.  Five 9’s of availability translates to a little more than 5 minutes of down time per year.  Making a design choice to sacrifice a huge amount of performance 364 days, 23 hours, 55 minutes a year in order to gain an advantage in the remaining five minutes a year doesn’t exactly jive with storage efficiency claims.  And if you think about it – an active/active dual-controller system operating with a failed controller will have exactly the same amount of system resources during those five minutes a year as the active/passive design – 50%!  So in fact you only gain the perception of maintaining performance because the performance level was half as great to start.

Active/passive is a 1980s design.  It made sense when processors were much slower, low-latency was limited to the backplane, and coding for HA was still in its infancy.  Now there are much better alternatives.  An N-way active design gives you full performance from all N controllers 99.999% of the time.  In the rare event of a controller failure or planned maintenance, service continues uninterrupted at (N-1)/N of original performance.Screen Shot 2013-12-29 at 8.53.30 AM

So which is the smarter design?  Do you want to design or buy a storage system that wastes 50% of its performance for 99.999% of the time just to keep the same 50% performance for an additional 5 minutes a year?  Or would you rather have the high performance of a linearly scaling cluster 99.999% of the time, with a manageable drop in performance the other 0.001% of the time (and remember the drop leaves you no worse off that the healthy performance of the active/passive array design)?

Imagine a car manufacturer sold you a car that has a V8, but it only ever runs on 4 cylinders and goes 50MPH – the other 4 cylinders you paid for don’t add any horsepower.  They don’t make the car go faster.  But if one of the active four cylinders has a failure, you’ve got some spares and can keep driving 50MPH.

Meanwhile, your friend bought a car that can have a V8 engine.  Or a V16.  Or a V32.  His car runs on all 32 cylinders and can go 400MPH.  If some cylinders fail he has to temporarily slow down to 350MPH – but this only happens 0.001% of the time.  Which car do you want to drive when you need to get someplace?  Even 350MPH is 7X faster than your car can ever go.

Of course, the 400MPH V32 car is XtremIO, except it has not 99.999%, but 99.9999% availability.  So it runs up to 8x faster 99.9999% of the time, and runs 7x faster during the other 0.0001%.  The race is won.  Hands down.

Other than the performance advantage, there are many other reasons why the scale-out, N-way active architecture of XtremIO is preferred over the dual-controller, active/passive approach.  First, is the ability to size your deployment based on I/O or capacity requirements.  With XtremIO, every X-Brick has two controllers and can perform 250K fully random 4K read IOPS, and 150K fully random 4K mixed read/write IOPS.  If a customer needs 300K mixed IOPS, he can choose a two X-Brick cluster.  If he has a workload requiring 600K mixed IOPS, he can configure a 4 X-Brick cluster.  This level of flexibility is not possible with any dual-controller architecture.

Another big advantage of XtremIO’s N-way active scale-out design is that there are never “stranded IOPS”.  Large capacity SATA disk drives suffered from “stranded capacity” – the drives got so big (but no faster) and eventually larger drives were useless because there wasn’t enough speed to move bulk data on and off the drives.  All-flash arrays have a similar problem.  There are so many IOPS available on the SSDs that the array controllers quickly become a bottleneck and leave IOPS “stranded” in the array.  But not with XtremIO.  Every 25 SSDs are balanced by two active controllers, allowing the collective system to deliver higher and higher levels of performance as capacity grows.

Customers often find this scale-out capability is exactly what they need when they size their database, server virtualization, and VDI infrastructures. We believe this is very important for all-flash arrays.  Performance is one of the primary reasons customers buy flash.  If you have a ceiling for performance while claiming you can grow capacity, there is a problem.  Some vendors will dismiss this by saying flash is so much faster than disk that performance is plenty even if bottle-necked by two controllers.  But history has shown that application developers quickly consume every bit of performance available to them and quickly ask for more.  What seems fast today can quickly be not enough tomorrow – especially as workloads begin to consolidate onto flash arrays.

If there are so many advantages to the XtremIO architecture over the dual-controller, active/passive approach, shouldn’t we expect every vendor to add this capability?  Most things are possible with enough time and money.  But architectural changes at this level are rare, disruptive, and expensive.  You always want to start with the right architecture from the very beginning.  We invested significant time and resources to do this with XtremIO.  It wasn’t the fastest path to market, but it did allow us to deliver a superior product.  It means you can count on superior performance all the time with XtremIO – even under worst-case failure conditions our N-way active scale-out model delivers the same performance as healthy dual-controller all-flash arrays.  The rest of the time it’s not even a close race.

By Robin Ren On December 26, 2013 ·

Leave a comment

Filed under Uncategorized