Virtual Geek

Solid State Disk will change the storage world…

You know when you hear your own company say something over, and over again – sometimes one can have their own doubts.

I remember the first time Donatelli (formerly at EMC, now at HP) told me: “in 2-3 years, there will be no place for a high-performance rotating rust disk, we’ll only have solid state and very large slow magnetic media” (that was about 1 year ago).

I was a bit skeptical.   

I was also talking to a customer about it yesterday, and we debated about cross-over points (not whether it would happen, but just timelines).

This just clears away any of my skepticism.

So – this article is on Anandtech – one of my favorite non-storage related IT sites.   If you don’t want to read the whole article, jump to this page.

Or – let me summarize it for you.   Read on if interested, or if after reading the article the “tada!” epiphany doesn’t jump right out at you.

Wednesday morning present – SRM 4 and Mirroview Insight for VMware

This is a FANTASTIC free tool that is included in the latest MirrorView SRA.

So – what does it do?

Well – you can run it before you setup a VMware Site Recovery Manager recovery plan, and it makes it much easier.   It will summarize whether the underlying replication pre-requisites have been met for individual VMs (some customer have VMs with VMDKs on multiple datastores, or VMs that have multiple RDMs), or entire datastores (some customers have spanned extents).   This is an incredibly powerful capability, and makes SRM+Replication setup much easier.

And… you can run it after you execute a Site Recovery Manager recovery plan, and use it to failback, including every step except sequencing and IP address changes.

Where do you get it?  It automatically installs when you download the MirrorView SRA from VMware here.

Here’s a great short demo of this awesome tool in action – check it out!

You can download a high-rez version of this demo here in WAV and MOV format.

Thanks to a couple people:

First of all – the incredible Giri Basava.   Part of CLARiiON engineering, he single-handedly developed the MirrorView Insight for VMware tool.   Giri is always awesome and coming up with great innovate ideas.   Above and beyond that one thing I admire about him in particular is amazing customer focus.   Whenever we’ve had a Mirrorview SRM case, bam – get it to Giri and it’s resolved.   Giri – you are incredible!

Second – in “Chad’s army” – Alex Tanner and Ben Dunning.   Alex – for his great work on this tool in beta – a lot of the things people will like are a result of Alex working hand in hand with Giri.  Alex has great “customers will dig this” instincts.   Ben - thanks for working on this demonstration video!

So, to recap for customers using EMC replication products and VMware:

• If you’re sitting on the fence about SRM – stop :-)   Join the more than 2000 SRM customers around the world – the majority of which are using EMC Storage Replication Adapters.  With SRM 4 adding vSphere 4 support, N:1 replication topologies, NFS datastores support, and faster recovery plan execution – there’s no reason to wait, particularly if you are virtualizing mission critical applications that need disaster recovery.   At EMC we’re happy to help you – regardless of your protocol, regardless of your scale, regardless of your recovery point objectives.  We can cover it all – heck, we can cover heterogenous (non EMC storage) use cases too!
• If you’re an EMC/VMware partner – make sure you check out the standard service kits available on Powerlink.  These can help you work with your customers to accelerate their VMware Site Recovery Manager deployments. 
• SRM 4 coverage: Celerra (both block and NAS), CLARiiON Mirrorview, and Recoverpoint SRAs have fully completed their VMware qual and are available.   In fact, I was happy to see that our engineering and qual work paid off – and Celerra and Mirrorview were there day one of the SRM 4 launch.
• The only EMC adapter who’s qual has not yet finished is SRDF, and it’s SRM 4 SRA is just about done.   Should be any day now.
• Each of the latest EMC remote replication adapters and replication products does something above and beyond the SRM basics at no extra cost.
  • EMC Celerra Replicator: has a vCenter plugin that automates failback for iSCSI (associating VM objects and making it simple to reverse replication, present back to the production site, renaming datastores, handling VMware un/registration, automating VM power-on activities)
  • EMC Mirrorview: has Mirrorview Insight for iSCSI, FC, and FCoE use cases.   This actually does even more than the Celerra failback tool – including changing VM networks.
  • EMC Recoverpoint itself directly integrates with vCenter, making mapping of dependencies and configuration a piece of cake, for iSCSI, FC, and FCoE use cases.

So.. What’s next on the topic of VMware-related DR?

• We couldn’t leave the Symm out with all this additional bonus stuff :-) The next SRDF SRA has tools to make configuration, setup, test, and failback much easier – this has been the #1 requested feature for SRDF use cases.
• The next version of the Celerra vCenter plugins expands the failback capability a fair amount.  While the Celerra SRA supports iSCSI and NFS, right now the failback capability is iSCSI only.  NFS failback support is around the corner.   Some additional failover handling (VMware network reconfig, start sequence) are all also being worked on.
• I’m corralling the product teams to make sure that the good ideas that they came up with independently get integrated across the product portfolios (I know this is what customers would like us to do).
• We’re heads down with VMware on the next major rev of SRM, which does a LOT.   Ultimately this will be the right long term vehicle for completely exhausive failback (and other powerful workflows!) management – and a lot of the lessons learnt from all this innovating extra tools we’ve developed are being applied.   Still early days down that project, but SRM will continue to expand it’s functionality!!

Oracle on vSphere on NFS - Rapid Test/Dev with EMC Replication Manager

A neat, short demonstration we did for a customer showing how they could leverage EMC Replication Manager for VMware-integrated snapshots of their NFS datastores for an Oracle use case.

In this scenario, EMC Replication Manager is doing a VMware-integrated NFS datastore (on an EMC Celerra) snapshot for a simplified, non-disruptive test/dev use case of an Oracle VM – while it’s doing about 6300 transactions/min generated by swingbench.   Like all EMC platforms, the Celerra can create simple, fast, writeable snapshots.   Like all EMC platforms (both block and NAS), it can be managed by EMC Replication Manager with VMware integration (correlation of objects with vCenter, and leveraging transient ESX snapshots as an intermediary step).

 

Download a high-rez version of the demo here.

BTW – you can use these technologies to make “repro in a physical” fast an easy.   The two primary ways are: 1) on block devices using physical mode RDMs, which can be quickly snapshotted and presented to a physical host; 2) on NAS devices, using dNFS into the guest allows you to simply mount on a physical host.   Simple, easy “get out of jail free” card for those virtualizing Oracle.  More on this later.

What’s coming next in this vein:

• very soon, the ability to do VM-level snapshot operations (in addition to datastore-level one shown here) on NFS datastores.   Today, datastore level snapshots are still useful for VM-level operations, but involves mounting the entire snapshotted datastore.   Restoring the VM to the original production datastore involves a “copy back” step which takes some time.  Still that’s much, much faster than a traditional restore.   I never think of array snapshots as a “complete backup solution”, but they are awesome for “why not restore that way if you have it available”.
• longer-term via the vStorage APIs for Array Integration or VAAI (which was demonstrated at VMworld 2009), similar hardware-offloaded VM-level replica capability on block devices.

Thanks Stephen Spellicy for the demo – you rock!

BTW – clever watchers of my team’s demonstrations (including this one) are finding hidden clues of other as of yet non-public Virtual Appliances we already have at our disposal as they are there in our vCenter inventories.   Someone found the Rainifinity GFV (still not public) in the one I recently posted (see the comment here).   See if you can find the unreleased EMC VM appliance in this demo :-)

Also – I’m not trying to be a tease :-)   If we haven’t made it public, it’s usually because there are still too many caveats (like need for specific devices, or VMDirectpath configurations that are narrow)…, but… working on it :-)

Monday morning present – EMC FMA Virtual Edition

UPDATE 5/22/2010: There is a newer FMA/VE available for play/demo/test here.

Here’s another Storage Virtual Appliance (joining the Celerra, Avamar, Atmos and many many others) – this time, EMC Rainfinity File Management Appliance – Virtual Edition (FMA/VE).  A big thank you to the EMC FMA team for doing the right thing and opening this up, in particular Scott Hall.  Thanks Scott!

What is EMC File Management Appliance?   It’s an appliance that can move files between filesystems transparently, enabling automated policy-driven tiering.  

For example, you can move files on a Celerra between Flash based tiers, FC based tiers, SATA based tiers, and in and out of deduped filesystems.    On the real product (not the current Virtual Edition – this is one difference between the virtual and physical versions…) you can also use it between Celerra and NetApp systems, and to and from NAS devices to Centera (gee – I wonder what could be the next cloudy tier? :-).

Why would you do it?   Most filesystems are mostly old stale data.  They consume all the storage costs as if they were actively used.   They are backed up every time a full filesystem backup is done.   For most customers – this can save a whackload of money, and make backups of any NAS environment work better.

If you’re interested, read on for details, downloads and documentation!

Cloud storage – what the hell is EMC building?

Just got back from 2 weeks in Europe with customers of every shape and size.  One thing that was clear to me was that we need to be more clear about our cloud storage strategy – as many of them were struggling with issues and didn’t know what we could do to help them.   Then again, I’m not sure that I will help make things more clear with this post - one thing I’m never is “talking point concise” :-)

So – what is our cloud storage strategy? 

First of all, it’s very important that cloud doesn’t imply “internet and public” by definition.   It does imply these things:

1. usage models (large aggregate pools, oversubscribed – not “hard provisioned/thick”)
2. consumption and economic models (pay-as-you go)
3. management models (zero-touch from a user perspective, with very “cross silo” tools from an administrative perspective).  

These core ideas and principles apply all infrastructure services (IaaS) models – including both compute and storage.   They also imply whether those services are constructed and maintained by an enterprise in within/across its walls as a purely internal service (presumably higher SLAs and security than public cloud), or by a service provide (presumably higher degrees of scale and oversubscription and therefore more efficient).

If this sounds “future speak” – it’s NOT.   EVERY customer I talked to was trying to build compute and storage self-service internal IT services.   They may call them different things, but when you describe it like the above and ask: “are you trying to do something like this?” the answer was always “YES!”.

As any reader of Chuck’s blog  (here, here, and here) knows – EMC’s view is that both Internal Clouds and External cloud models are winners for enterprise customers and service providers, and once we crack the nut of open-inter cloud standards and federation – whole new technology and economic vistas open up for our customers.   Even before that – well – if you’re building a new application/infrastructure/datacenter – customers tend to be thinking on these lines.

But… I leave the high-level thinking to Chuck (he does a much better job of describing this at a high level than I do), I want to spell out the implications on this in perhaps a lower level of detail.    I also sketched it out for a couple customers, and thought the diagrams are perhaps useful to make this simple to comprehend.

If you’re interested – read on…

EMC Platform/vSphere Best Practices Webcasts

Folks – these webcasts are very popular.   Some of you missed the Symmetrix/V-Max one, and wondered if it was recorded.  They are all recorded, and just take a few days to get posted.

You can see the recording of the Symmetrix/V-Max one here:

http://www.emc.com/events/2009/q4/10-01-09-oct-vmware.htm

The CLARiiON/vSphere one was on the 8th, and should be posted soon.

The next up to bat is the Celerra webcast (Oct 15th 11-12 EST) which will cover best practices in NFS, iSCSI and FC use cases, latest testing results, how to leverage snapshots and replication as easily as possible.   Remember that whatever you hear/learn, you can try for yourself using iSCSI/NFS on the Celerra VSA which is openly available here. 

Register for the Celerra webcast here:

http://www.emc.com/events/2009/q4/10-15-09-vmware-oct.htm

For those of you that don’t see the pattern, Thursday at 11-12 is “EMC/VMware webcast thursday” – every week, we’ll have a topic.   You can see upcoming topics here:  http://www.emc.com/events/events-calendar.esp  (you can filter on Live Webcasts).   Upcoming topics include vSphere management integration, and also the updated validation results of Oracle on vSphere (both performance, as well as test/dev recovery models and more).

Feedback welcome – my guidance to the team is always firm on these guidelines, so tell me if they are solid – I make most of the webcasts personally, but not all of them….

• Keep them germane
• Keep them Technical
• Demonstrate, don’t powerpoint
• Keep marketing to a minimum

Hold us to that – and we’ll try our best to make it a good use of your time – let me know whether you find them useful!

ix2-200d – are you a REAL cheapskate looking for iSCSI/NAS storage?

The Iomega ix2-200d is now out – this is the 2 drive next-gen iomega platform – basically similar features to the ix4, but in a very small 2 drive form factor.  iSCSI, NFS, ships in 1,2,4 TB raw configurations (RAID 1/JBOD). 

At home, I use my ix4 to stream movies and my itunes library to my PS3 (which does an awesome job with divx) and Xbox360.   The other thing that’s awesome is the systems support drive spin-down, which is a nice power-reducing effect.

Here’s info in a nice little youtube-sized nugget…

Remember folks – its no rocket.   As I wrote in Chapter 6 in the Mastering vSphere 4 book – storage newbies think only in terms of capacity.   IOps and throughput matter.  If you only have 2 5400 RPM drives – well – even though you COULD put a ton of VMs on there, just don’t expect it to be a rocket.

On the other hand – it does start for $297 :-)

When I get back from this europe road trip – will benchmark my ix4 and ix2.

Coolest thing I’ve seen all week….

This is definitely the coolest thing I’ve seen all week (and I’ve had a crazy week – Toronto, Munich, London, Madrid).

Worth checking out.   I’ve been playing with the same bits – but not with that rig :-)

http://www.youtube.com/watch?gl=US&v=APVP0DUu34c

Cloud talk, and cloud action.

So, EMC has had our Cloud storage solution up and running for some time, and petabytes are flying out the door (good!) to service providers.  This is object-based storage, exabyte scale, with a global object distribution, federation and global namespace.   The magic is in the software – as it  runs on what is in essence low cost, dense JBOD – turning it into a global cloud storage model, with the economic model to match.

You access it generally via a Web 2.0 interface, it has APIs to program to.   if you know Amazon S3, it’s a real live Amazon S3 alternative.   We each do something a little different, so people can make their own choices about what they like.   Think of this as the architectural model if you were building your own Flickr, or YouTube.   You would NEVER built those on a traditional array model (the economics are whacky).

You can sign up for it here.

What’s next up to plate (and live!)?   Well - Cloud compute services.    This adds compute resources, a layer of virtualization (provided by VMware of course) and self-provisioning automation.

In fact those operational characteristics (self-provisioning, elasticity, and pay-as-you go) are the definition of cloud more than anything else (and why the word “cloud” can apply to a virtualized datacenter with those additional characteristics).

This is now up and running, and you can sign up for that here.   Cloud compute (hate to draw the obvious analogy, but Atmos Online Compute Service competes with Amazon EC2) can be used in general with varying workloads – some can use object storage models (can deliver storage at an order of magnitude lower cost), some need traditional scale-out enterprise storage architectures (which can be thinly provisioned, going wide across the array, and hit the economic models used today for $/GB/month for cloud compute scenarioes – but is still more expensive by several factors than the cloud object storage models.

This, by the way is why most vCloud launch partners are using both EMC V-Max and Atmos.   Both of these are mature, and have been shipping for a while.

Our strategy here is clear.

Are we trying to compete with our service provider partners?   HECK NO. 

EMC will not build global datacenters (Atmos Online and Atmos Cloud Compute Services operate out of two US-based datacenters).  So why are we doing it?   To help accelerate our partners.  Getting real time under our belt, not just with the technology, but the APPLICATION of the technology.   This also lets us innovate around them faster, as well as bring the intellectual property to help service providers who want to get into this business.

Pretty darn cool stuff.    Remember – this is not science future – it’s science fact.  

What’s science future – and this I’m close to – is how we will start to link common metadata models in both extremes, broaden the continuum (right now there are two extremes in SLA and cost models), and start live federation between sites in the cloud compute use cases.  

I think I’ve got the coolest job in the world!

Customer Council – London…

One of my favorite work-related tasks are customer councils.   For those of you who aren’t EMC customers (though I’m sure other vendors do something similar), these are intimate off-site with our customers.  Sales are not allowed.   all topics are on the table – what we do well and not well – what they like, and what they don’t.   We share roadmaps and plans - and customers give us feedback.

Just finished one here in London (in this example – 12 large UK customers) – focused on what we’re doing re VMware and EMC, and also in the VMware/Cisco/EMC efforts.  Fascinating dialog inlcuding prioritization input over the next 6, 12, and 24 month periods.   Fascinating input.

To the customers who attended – thank you!

On to the IP expo….