Storagebod Rotating Header Image

Data Centres

Hursley Heroes….

It has been very gratifying to see the V7000 and SVC getting some real love from IBM over the past couple of years; it’s a product which is going from strength to strength. The latest bunch of announcements are good; bigger, faster and yet presented with realism; Barry is not one to try and pull the wool over our eyes by quoting nonsensical performance figures.

The two features that catch the eye are the compression capabilities that should work very well for databases and the likes; Storage Pool Balancing, ensure that all disks in a pool are delivering the same performance is the feature which is most important.

I wonder if the Storage Pool Balancing is the first step to putting a disk-hospital into the V7000; allowing a much lower-touch maintenance and a subsequent reduction in maintenance costs (to IBM, not to the customer obviously). Identifying disks that are performing slower than their peers even minuscule levels is often the first indication that something is going wrong; it is also very important in IBM’s declustered RAID…something not yet announced in the V7000/SVC range but must be expected as disks get ever bigger.

The use of Intel’s QuickAssist technology to enable the compression chip should bring future enhancements such as encryption and deduplication.

So the future is pretty bright for the SVC and V7000…as long as IBM can actually manage to market and sell it.

If you also factor in the common look and feel across the storage range; you’d almost believe that IBM have a real future in storage…

p.s I see that IBM are persisting with the V7000U as well; I keep wondering whether the next V7000U announcement will be a withdrawal from market but I guess IBM need to keep trying to sell a traditional filer.



Not So Potty

Virtual Openness

I don’t always agree with Trevor Pott but this piece on ServerSAN, VSAN and storage acceleration is spot on; the question about VSAN running in the kernel and the advantages that brings to performance; and indeed, I’ve also heard comments about reliability, support and the likes over competing products is very much one which has left me scratching my head and feeling very irritated.

If running VSAN in the kernel is so much better and it almost feels that it should be; it kind of asks another question, perhaps I would be better running all my workloads on bare-metal or as close as I can.

Or perhaps VMware need to be allowing a lot more access to the kernel or a pluggable architecture that allows various infrastructure services to run at that level. There are a number of vendors that would welcome that move and it might actually hasten the adoption of VMware yet further or at least take out some of the more entrenched resistance around it.

I do hope more competition in the virtualisation space will bring more openness to the VMware hypervisor stack.

And it does seem that we are beginning towards data-centres which host competing virtualisation technologies; so it would be good if that at a certain level that these became more infrastructure agnostic. From a purely selfish point of view; it would be good to have the same technology to present storage space to VMware, Hyper-V, KVM and anything else.

I would like to easily share data between systems that run on different technologies and hypervisors; if I use VSAN, I can’t do this without putting in some other technology on top.

Perhaps VMware don’t really want me to have more than one hypervisor in my data-centre; the same way that EMC would prefer that all my storage was from them…but they have begun to learn to live with reality and perhaps they need to encourage VMware to live in the real world as well.  I certainly have use-cases that utilise bare-metal for some specific tasks but that data does find its way into virtualised environments.

Speedy Storage

There are many products that promise to speed-up your centralised storage and they work very well, especially in simple use-cases. Trevor calls this Centralised Storage Acceleration (CSA); some are software products, some come with hardware devices and some are mixture of both.

They can have some significant impact on the performance of your workloads; databases can benefit from them especially (most databases benefit more with decent DBAs and developers how-ever); they are a quick fix for many performance issues and remove that bottleneck which is spinning rust.

But as soon as you start to add complexity; clustering, availability and moving beyond a basic write-cache functionality…they stop being a quick-fix and become yet another system to go wrong and manage.

Fairly soon; that CSA becomes something a lot closer to a ServerSAN and you are sticking that in front of your expensive SAN infrastructure.

The one place that a CSA becomes interesting is as Cloud Storage Acceleration; a small amount of flash storage on server but with the bulk of data sitting in a cloud of some sort.

So what is going on?

It is unusual to have such a number of competing deployment models for infrastructure; in storage, we have an increasing number of deployment models.

  • Centralised Storage – the traditional NAS and SAN devices
  • Direct Attached Storage – Local disk with the application layer doing all the replication and other data management services
  • Distributed Storage – Server-SAN; think VSAN and competitors

And we can layer an acceleration infrastructure on top of those; this acceleration infrastructure could be local to the server or perhaps an appliance sitting in the ‘network’.

All of these have use-cases and the answer may well be that to run a ‘large’ infrastructure; you need a mixture of them all?

Storage was supposed to get simple and we were supposed to focus on the data and providing data services. I think people forgot that just calling something a service didn’t make it simple and the problems go away.


Licensed To Bill

‘*sigh* Another change to a licensing model and you can bet it’s not going to work out any cheaper for me’ was the first thought that flickered through my mind during a presentation about  GPFS 4.1 at the GPFS UG meeting in London (if you are a GPFS user in the UK, you should attend this next time…probably the best UG meeting I’ve been at for a long time).

This started up another train of thought; in this new world of Software Defined Storage, how should the software be licensed? And how should the value be reflected?

Should we be moving to a capacity based model?

Should I get charged per terabyte of storage being ‘managed’?

Or perhaps per server that has this software defined storage presented to it?

Perhaps per socket? Per core?

But this might not work well if I’m running at hyperscale?

And if I fully embrace a programmatic provisioning model that dynamically changes the storage configuration…does any model make any sense apart from some kind of flat-fee, all-you-can-eat model.

Chatting to a few people; it seems that no-one really has any idea what the licensing model should look like. Funnily enough; it is this sort of thing which could really de-rail ServerSAN and Software Defined Storage; it’s not going to be a technical challenge but if the licensing model gets too complex, hard to manage and generally too costly, it is going to fail.

Of course inevitably someone is going to pop-up and mention Open-Source…and I will simply point out, RedHat make quite a lot of money out of Open-Source; you pay for support based on some kind of model. Cost of acquisition is just a part of IT infrastructure spend.

So what is a reasonable price? Anyone?



Too Cheap To Manage…

Five years or so when I started this blog, I spent much time venting my spleen at EMC and especially the abomination that was Control-Center; a product so poor that a peer in the inudstry once described it as being too expensive even if it was free.

And yet still the search for the perfect storage management product still continues; there have been contenders along the way and yet they still continue to fall short and as the administration tools have got better and easier to use, the actual management tools have still fallen some way short of the mark.

But something else has happened and it was only a chance conversation today that highlighted this to me; the tenuous business case that many have been purchased on has collapsed…many storage management products are purchased with the business case that ultimately that they will save you money by allowing you to right-size your storage estate….they will maximise the usage of the estate that you have on the floor.

Unfortunately and it surprises to say this; the price of enterprise storage has collapsed…seriously, although it is still obviously too expensive (I have to say that); the price of storage management products has not declined at the same rate. This means that it is doubtful that I can actually save enough capacity to make it worth my time trying too hard and putting in a tool to do so, the economics don’t actually stack up.

So there has to be whole new business case around risk mitigation, change-planning, improved agility…or the licensing model that tends to be capacity-based in some form or another has to be reviewed.

Do we still need good storage management tools? Yes but they need to focused on automation and service delivery; not on simply improving the utilisation of the estate.

Thin-provisioning, deduplication, compression and the likes are already driving down these costs; they do this ways that are easier than reclaiming orphaned storage and even under-utilised SAN ports.And as long as I am clever, I can pick-up a lot of orphaned storage on refresh.

If ‘Server-SAN’ is a real thing; these tools are going to converge into the general management tools, giving me a whole new topic to vent at..because most of these aren’t especially great either.

p.s If you want to embarrass EMC and make them sheepish…just mention Control-Center…you’d think it’d killed someone..

All The Gear

IBM are a great technology company; they truly are great at technology and so many of the technologies we take for granted can be traced to back to them. And many of today’s implementations still are poorer than the original implementations.

And yet IBM are not the dominant force that they once were; an organisational behemoth, riven with politics and fiefdoms doesn’t always lend itself to agility in the market and often leads to products that are undercooked and have a bit of a ‘soggy bottom’.

I’ve been researching the GSS offering from IBM, GPFS Storage Server; as regular readers of this blog will know, I’m a big fan of GPFS and have a fair amount installed. But don’t think that I’m blinkered to some of the complexities around GPFS; yet it deserves a fair crack of the whip.

There’s a lot to like about GSS; it builds on the solid foundations of GPFS and brings a couple of excellent new features into play.

GPFS Native RAID; also known as declustered RAID is a software implementation of micro-RAID; RAID is done at a block level as opposed to a disk level; this generally means that the cost of rebuilds can be reduced and the time to get back to a protected level can be shortened. As disks continue to get larger, conventional RAID implementations struggle and you can be looking at hours if not days to get back to a protected state.

Disk Hospital; by constantly monitoring the health of the individual disks and collecting metrics for them; the GSS can detect failing disks very early on but there is a dirty secret in the storage world; most disk failures in a storage array are not really failures and could be simply recovered from, a simple power-cycle can be enough or a firmware reflash can be enough to prevent a failure and going into a recovery scenario.

X-IO have been advocating this for a long time; this can reduce maintenance windows and prevent unnecessary rebuilds. It should reduce maintenance costs as well.

Both of these technologies are great and very important to a scalable storage environment.

So why aren’t IBM pushing GSS in general; it’s stuffed full of technology and useful stuff?

The problem is GPFS…GPFS is currently too complicated for many, it’s never going to be a general purpose file system. The licensing model alone precludes that; so if you want to utilise it with a whole bunch of clients, you are going to be rolling your own NFS/SMB 3.0 gateway. Been there, done that…still doing that but it’s not really a sensible option for many.

If IBM really want the GSS to be a success; they need a scaleable and supported NAS gateway in front of it; it needs to be simple to manage. It needs integration with the various virtualisation platforms and they need to simplify the GPFS license model…when I say simplify, I mean get rid of the client license cost.

I want to like product and not just love the technology.

Until then…IBM have got all the gear and no idea….

A two question RFP….

Is it easy?

Is is cheap?

Pretty much these are the only two questions which interest me when talking to a vendor these days; after years of worrying about technology, it has all boiled down to those two questions. Of course, if I was to produce an RFx document with simply those two questions, I’d probably be out of a job fairly swiftly.

But those two questions are not really that simple to answer for many vendors.

Is it easy? How simply can I get your product to meet my requirements and business need? My business need may be to provide massive capacity; it could be to support many thousands of VMs, it could be to provide sub-millisecond latency.  This all needs to be simple.

It doesn’t matter if you provide me with the richest feature-set, simplest GUI or backwards compatibility with the ENIAC  if it is going to take a cast of thousands to do this. Yet still vendors struggle to answer the questions posed and you often get the response to a question you didn’t ask but the vendor wants to answer.

Is it cheap? This question is even more complicated as the vendor likes to try to hide all kinds of things but I can tell you; if you are not upfront with your costs and you start to present me with surprises, this is not good.

Of course features like deduplication and compression mean that the capacity costs are even more opaque but we are beginning to head towards the idea that capacity is free; performance costs. But as capacity becomes cheaper, the real value of primary storage dedupe and compression for your non-active set that sits on SATA and the likes begins to diminish.

So just make it easy, just make it cheap and make my costs predictable.

Be honest, be up-front and answer the damn questions….

Drowning in Roadmaps…

Roadmap after roadmap at the moment; bring out your roadmaps. Of course, this causes me a problem as I’ve now seen roadmaps going way off into the future and it is a pain because as soon as I start speculating about the future of storage; people seem to get very worried about breach of NDAs.

But some general themes are beginning to appear

1) Traditional RAID5 and RAID6 data protection schemas are still in general the go to for most of the major vendors…but all are acknowledging there are problems and are roadmapping different ways of protecting against data loss in the event of drive failures. XIV were right in that you need as many drives as possible taking part in the rebuild; they may have been wrong with specifics.

2)Every vendor is struggling with the appliance versus software model. It is almost painful to watch the thought-processes and the conflict. Few are willing to take the leap into a pure software model and yet they all want to talk about Software Defined Storage. There are some practical considerations but it is mostly dogma and politics.

3)Still the discussions about running workloads on storage arrays directly seem to rage and with little real clue as to what, how and why you would do so. There are some workloads that you might but the use-cases are not as compelling as you might think.

4)Automated Storage Tiering, it appears to be getting better but it still seems that people do not yet trust it fully and are wasting a huge amount of cycles second guessing the automation. Most vendors are struggling with where to go next.

5)Vendors still seem to be overly focussed on building features into general purpose arrays to meet the corner-cases. VDI and Big Data related features pepper roadmaps but with little comprehension of the real demand and requirement.

6)Intel have won the storage market or at least x86 has. And it is making it increasingly hard for vendors to distinguish between generations of their storage…the current generations of x86 could well power storage arrays way into the future.

7)FCoE still seems to be more discussed than implemented; a tick-box feature that currently outside some markets has no demand. 16 Gig Fibre-channel is certainly beginning to appear on the concrete side of the roadmaps; I’ve seen 40GbE on a couple now.

8)Flexibility of packaging and physical deployment options is actually a feature; vendors are more willing to allow you to re-rack their kit to fit your environment and data-centre.

9)The new boys on the block feel a lot like the old boys on the block…mostly because they are.

10)Block and File storage are still very resilient against the putative assaults of Object Storage.

11)The most compelling feature for many of us at the high-end is the procurement model that moves us to linear pricing. There are still struggles how to make this happen.

And yet expect big announcements with marketing splashes in May…Expect more marketing than ever!!!


So you’ve founded a new storage business; you’ve got a great idea and you want to disrupt the market? Good for you…but you want to maintain the same-old margins as the old crew?

So you build it around commodity hardware; you use the same commodity hardware as I can buy off the shelf; basically the same disks that I can buy off the shelf from PC World or order from my preferred Enterprise tin-shifter.

You tell me that you are lean and mean? You don’t have huge sales overheads, no huge marketing budget and no legacy code to maintain?

You tell me that it’s all about the software but you still want to clothe it in hardware.

And then you tell me it’s cheaper than the stuff that I buy from my current vendor? How much cheaper? 20%, 30%, 40%, 50%??

Then I do the calculations; your cost base and your BoM is much lower and you are actually making more money per terabyte than the big old company that you used to work for?

But hey, I’m still saving money, so that’s okay….

Of course, then I dig a bit more…I want support? Your support organisation is tiny; I do my due diligence,  can you really hit your response times?

But you’ve got a really great feature? How great? I’ve not seen a single vendor come up with a feature that is so awesome and so unique that no-one manages to copy it…few which aren’t in a lab somewhere.

In a race to the bottom; you are still too greedy. You still believe that customers are stupid and will accept being ripped off.

If you were truly disruptive….you’d work out a way of articulating the value of your software without clothing it in hardware. You’d work with me on getting it onto commodity hardware and no I’m not talking about some no-name white-box; you’d work with me on getting it onto my preferred vendor’s kit; be it HP, Dell, Lenovo, Oracle or whoever else…

For hardware issues; I could utilise the economies of scale and the leverage I have with my tin-shifter; you wouldn’t have to set-up a maintenance function or sub-contract it to some third party who will inevitably let us both down.

And for software support; well you could concentrate on those…

You’d help me be truly disruptive…and ultimately we’d both be successful…

Storage Blues…

January not even out yet and already we have an interesting technology market happening; IBM’s withdrawal from the x86 server market does lead to a number of questions. Both on the future of IBM but also on what IBM feel the future of the market is; yet could this be another market that they withdraw from only to long-term regret as they did with the network market allowing Cisco to dominate?

IBM’s piecemeal withdrawal from the hardware market; a retreat to the highlands of the legacy enterprise market in hardware will lead to questions across the board as to what the future is for any IBM hardware. I am not sure of the market acceptance of their converged compute/network/storage strategy in the form of PureSystems; their me-too ‘Block’ offering but surely this is dead-duck now; Lenovo may continue to make the x86 components for IBM but how committed can we feel that IBM is to this. IBM appear to have completely ceded this space to their competitors; personally I’m not convinced by most of the converged offerings and the value but to completely cede a market seems to be rash.

But how does this impact IBM storage?

The heart of IBM’s Storwize product set is x86-based servers; SVC especially was ‘just’ an IBM server. IBM were one of the first companies who really leveraged the idea of the server as storage; Shark is and was simply a pair of RS/6000 or pSeries boxes, this has allowed them to utilise and share R&D across divisions. Something which should have been an advantage and enabled them to do some clever stuff; this stuff  they demonstrated yet never delivered.

Now there is no reason for them to simply source the servers from others, the same as almost every other storage company in the world and it moves the Storwize product set firmly into the realms of software (it was anyway) but will IBM move Storwize to a software-only product?

There is part of me who really feels that this is inevitable, it may be as a reaction to a move by a competitor; it may be as a move to enable a vV7000 to run as a cloud appliance? It may well end up being the only way that IBM can maintain any kind of foothold in the storage market.

No I haven’t forgotten XIV or IBM’s Flash offerings; XIV is a solid Tier 1.5 offering but it is also a collection of servers. XIV’s issue is really scalability and simply putting larger drives in is just reducing the IOP density. The Flash offering is as good as many and if you want raw performance without features; it is worth considering.

IBM’s GSS could be built into something which scales and with many of the ‘features’ of XIV. And in a software only IBM Storage strategy; it could develop into a solid product if some of the dependency on specific disk controllers could be relaxed. Yet the question has to be whether IBM has time.

And yet without either a scalable NAS or Object store; IBM have some real problems. None of which are are really hardware problems but moving away from building your base platform probably makes none of them easier to solve.

Or perhaps if they concentrate on software and services….

Already Getting Busy…

I’ve not been away but a mixture of illness, Christmas and general lethargy have meant that I’ve not bothered with writing for a bit. But 2014 and a new year appears to be upon us and I do wonder what it is going to bring us, especially in the world of IT infrastructure.

As we ended 2013, we saw both winners and losers in the world of Flash for example; Violin crashing as they struggle to increase sales and reduce burn; yet Pure seem to be on a stellar rise and hiring like maniacs. A UK launch is imminent and they are going to be interesting to watch. All Flash Arrays are still very much niche and even companies who need them are holding off on making any big decisions.

I’ve already spoken to a hybrid vendor this year; pushing their hybrid is good enough for most cases, very tied to the virtualisation use-case. And yes, VDI all over their powerpoints as a use-case. 2014, the year when VDI happens!!

I expect that I’ll spend time with more hybrid vendors who are playing some kind of chicken with SSD/Disk ratios; how low can they go? However, I’m also seeing more KVM/Openstack appearing on road-maps as they begin to realise that VMware might not be the only game in town.

I’m sure we’ll see more hype around hyper-convergence as attempts continue to build a new mainframe and I shall continue to struggle to work out why anyone wants to? I like being able to scale my infrastructure in right place; I don’t want to have to increase my compute to increase my storage and vice versa. Flexibility around compute/storage and network ratios is important.

Yet convergence of storage and compute will continue and there’s potentially some real challenge to the traditional storage technologies there. If I was building a new infrastructure today, I’d be looking hard whether I needed a SAN at all. But I wouldn’t be going straight to a hyper-converged infrastructure; there be dragons there I suspect.

I’ve already had my first vendor conversation where I’ve suggested that they are actually selling a software product and perhaps they should drop the hardware part; that and asking why the hell were they touting their own REST API for cloud-like storage…if industry giants like EMC have struggled against the Amazon juggernaut, what makes they think that they are any different?

And marketing as differentiation will probably continue….especially as the traditional vendors get more defensive around their legacy products.  No-one should get rich selling disk any more but it won’t stop them all trying.