Storagebod Rotating Header Image

Storage

Complex is the new Simple

Feb 7th, 2012
by Martin Glassborow.
1 comment

As EMC add yet another storage array to their product line-up in the form of Project Thunder or as I like to call it, The Thunderbox; questions are going to be asked about product confusion and clash. VMAX, VNX, vPlex, Isilon, Atmos and now Thunder form a product line which is formidable but may lack a certain amount of clarity.

But is this the problem that many of us think,maybe we need a different way of thinking? If you look at most workloads, I would say that about 80% of them could be met by a single product but the problem is which 80% and can the other 20% be met by another product? This would at least keep it down to two products.

However my experience is beginning to suggest otherwise and although a great majority of our workload can be met by a single solution; we have some some niche requirements which can only be met by specific tools. There is some irony that one of my niche requirements is actually general purpose IT storage but that’s the industry I work in; your niche requirements will probably be different but there is no point trying to hammer those requirements onto devices which will do a slightly less than adequate job at best.

At the moment, we manage over a dozen different storage technologies; granted some of them do overlap and some of them are only there because of slightly dubious requirements from application vendors but we don’t stress about a new technology coming in and supporting it. The principles for management are common and once you can manage one technology, you can pretty much manage them all.

Our job would be a lot harder if we tried to force applications onto a common platform; so despite appearances from the outside looking in, our platform’s complexity has actually ensured that our jobs are simpler.

What vendors could do and some have started to do is to ensure that their own product families have common interfaces with converged look and feel. IBM for example have made great strides in this and it is one thing that EMC could take away from IBM.

But a rich product set does not have to be complex although it does need to be explained so that customers understand the use case.

Cache Splash

Feb 6th, 2012
by Martin Glassborow.
4 comments

It’s funny, I had a brief discussion about my blog with my IT director at work today; I wasn’t aware that he was aware that I blogged but it seems a couple of people outside of work had outed me, in what appears to be very complementary terms; he was pretty relaxed about my blog and one of his comments was that he assumed I discussed new products and I said I did.

But on the way home, I thought about it and to be quite frank, I used to talk a lot about new products but I don’t really do so these days. So it is ironic that today, I’m going to knock out a quick blog about EMC’s VFCache announcement; they don’t need the publicity but I’m going to talk about it anyway.

VFCache is very much a version 1.0 product from what I can see; EMC appeared to have set their bar quite low in what they are trying to achieve with this release; it appears that they’ve very much targeted Fusion-IO pretty much directly and decided to go after them from the get-go. Trash them early and don’t let another NetApp happen.

Go for engineering simplicity and don’t fill the product full of features….yet! Keeping it simple means that EMC can accelerate any array, not just an EMC array but in the future when new features come along, many of these might well only be available with an EMC back-end array. You’ve bought your Flash card, if you really want value….you need to partner it with an EMC array.

And in fact, to really leverage any server-side flash product; you probably do need array-awareness to ensure that you don’t do economically silly things like storing multiple copies of the same information in different caches; how many times do you want to cache the same data.

You need an efficient way of telling the array, ‘Oi I’ve cached this, you don’t need to’; this will allow you to utilise the array cache for workloads which might not easily support server-side caching currently. Perhaps at some point we’ll see a standard but standards are rarely fast moving in storage.

I also expect to see EMC build in some intelligence to allow it to leverage the split card capability; perhaps using PowerPath to flag that actually you might want to consider using the split card capability to gain performance?

I’d also be interested in seeing advancing modelling tools which allowed you to identify those servers and workloads which would most benefit from VFCache and what the impact is on the other workloads in the data-centre. If you accelerate one workload with VFCache and hence free up cache on the shared-array, do all workloads benefit? Can I target the deployment at key servers?

Deduplication is coming but it needs to be not at the expense of latency.

And of course there is the whole cluster-awareness and cache-consistency thing to sort out and perhaps this whole thing is a cul-de-sac whilst we move to flash-only-shared-storage-arrays…that’s until the next super-fast technology comes along.

Yes, EMC’s annoucement is very product 1.0 and a bit ‘ho-hum’ but the future is more interesting. Storage, Snorage? Sometimes but it’s impact sometimes wakes you up with a bit of a shudder’.

I wonder who is going to announce next or what the next announcement might be. 2012 might be a bit more interesting.

 

Refactoring the Future….

Feb 5th, 2012
by Martin Glassborow.
1 comment

Most of my career has been spent in infrastructure and I guess infrastructure related technologies is the thing which really interests me professionally but I have spent time working as a developer, leading a development team managing the transition to agile development and more recently, as well as running a storage team, I also manage an Integration and Test team; this has hopefully led me into becoming more rounded professionally. And I think we as Infrastructure techies have a lot to learn from our colleagues in development; they’ve probably got more to learn from us but we can learn something from them.

Recently I’ve been wondering if Infrastructure could be treated more as an application; something that is more dynamic and less static than it is traditionally. Can an Infrastructure be Agile? And can it be managed in an Agile manner? The received wisdom is almost certainly, no! More outages are caused by change in infrastructure than any other cause;’  ‘if aint broke, don’t fix it’ is not actually a bad way to go but is that necessarily the case?

One of the big changes that Cloud brings should be that applications are engineered to cater for failure and change; an application should be able to move, replicate, grow, shrink and be resilient in spite of infrastructure.

Now this could mean that we could deploy, change and improve infrastructure a lot more rapidly; refresh becomes a constant task and not something which needs a special project to do, tuning and maintenance become routine and not scary. We have an infrastructure that is being constantly refactored in the same way that good developers refactor code; we fix kludges and workarounds, no longer living with them because we are fearful of the consequences of changing them.

Yet we also need discipline not to change for change’s sake; anyone who has run an Agile team will tell you about the importance of eg0-less developers (an oxymoron if there ever was) but you can run into the problem where more time is spent refactoring than adding value. I am not convinced that we will have this problem initially in Infrastructure, the culture of  ‘if aint broke, don’t fix it’ is heavily ingrained.

Refactoring is a small part of Agile but it is something that we can learn from in our Infrastructure domains. However there is a big problem; vendors don’t really like change and they throw all kinds of obstacles in your way, like interoperability matrices and the like. They really don’t like the idea of incremental change; they want big change where they can bill big tickets and big services.

I think that vendors and service providers who are going to win big are those who can deal with incremental change and improvements; those who are reliant on big changes and refresh cycles may well struggle as we move to a more dynamic model. I include internal IT suppliers in  this as well; the big change is a cultural change to accept that change is constant and good.

Your Data, Your Responsibility

Jan 24th, 2012
by Martin Glassborow.
No comments yet

I’ve been thinking recently about the post-PC era and what it really means; for some people it means the end of the desktop and the traditional PC but I think that this is slightly wrong-headed. For me, the post-PC era is my content anywhere and at any time.

Access to data is more important than anything but you might still use a traditional desktop to do heavy-lifting and manipulation; for example, tablets are great but for many tasks, I would still want to use a traditional keyboard, mouse and big screen. But when I’m away from base, I still want access to my data and perhaps do some lightweight manipulation.

So the post-PC world is moving us away from a single tethered end-point device to a multitude of devices, some mobile and some fixed. The applications we use on these devices may be different, in both scope and function but the data will be common and accessible everywhere.

This will bring challenges to us as individuals and as businesses; where do we store that data and how do we protect that data, both ensuring it is stored securely but also that it remains available. The recent Megaupload closure has already lead some people to question the long-term viability of cloud-storage. What happens if the site you store your data on is suddenly shut-down?

Question where you are putting your data; if it becomes obvious that a site has a slightly dubious reputation, then perhaps you should ask yourself whether you want to rely on its availability. But even if it is a site which has the highest reputation, ‘Shit Does Happen’; so you probably want to ensure that you have multiple copies stored in multiple places.

But also be aware of the underlying service, if both your Cloud storage providers are reselling storage from the same Cloud provider; question again.

Your data, your responsibility…

 

Meaningless Metrics!

Jan 20th, 2012
by Martin Glassborow.
3 comments

Recently we’ve had a bit of a senior management reshuffle and as a result, people are in ;prove that their teams are valuable’ mode again. Inevitably this means producing metrics to show why your team is brilliant, so my manager came along and asked me to produce some metrics about our storage estate.

So off I went and produced some pretty graphs showing storage growth, increase in number of technologies supported and some other things that I thought might show why my team does an excellent job. One of the problems with running a very good team is that they tend to have a relatively low visibility; they don’t cause problems and things don’t often break. Most of the time, people don’t know that we are here.

Anyway, as is the way of these things; the usual comment comes back; how many terabytes per head do we manage and what is the industry average?  Now with over two petabytes of active data per head under management, I could claim that my team is great but to be honest no-one really knows what the industry average is and would it be meaningful anyway? I’ve seen from 50Tb to 1 Petabyte quoted but with a figure of 150-200Tb most oft quoted; so my team could be good, great or downright amazing (it’s somewhere between the last two).

However, this is all meaningless and becomes more meaningless the more that industry changes. For example, we are managing what is closer to a big-data environment; big data environments have large infrastructures but if I am being honest, they are not especially hard to manage.

We rely on a fair amount of automation and standardisation; applications often do a lot of the storage management function and although the storage infrastructure grows, it tends not to change massively. Allocations are large but they tend to be relatively static; in that once allocated, it does not move around a lot; we make a lot of use of clustered file-systems and most of the work we do is non-disruptive. We add nodes in and even if node fails, it tends not to take every thing with it; we can live with a node down for weeks; the applications are resilient and services can generally cope with failures.

We have our busy times but it does generally run pretty smoothly; most of our time is spent on working out how we can make it run even more smoothly and how we improve the service, which in my opinion is exactly how it should be. The best support teams look busy but not stressed; hero cultures are not where it’s at.

So I’ve given my boss a figure but I am really not sure that it has a lot of value. Lies, Damn Lies and Metrics!

Dell of Dilemma

Jan 12th, 2012
by Martin Glassborow.
5 comments

I have just spent time at the Dell Storage Forum event 2012 (disclosure: Dell kindly paid for the hotel and subsistence) and although is no-where near the size of an EMC-World type event, everyone seemed to be having a good time and there were lots of smiling attendees.

Compellent customers seem to be almost uniquely happy with their supplier and this does not appear to have changed too much even after year of Dell’s stewardship; so it appears that Dell are at least not messing this up.

But I think Dell have a dilemma, how do you meld a set of acquired IP and product into a strategy? Especially with all due respect to Dell, in a company which does not have a massive history in product innovation.

Dell have been a great client-side supplier with strong supply-change management and delivery but they are not the company you immediately think about when talking about innovation.

Many of the announcements made at the Storage Forum are arguably just catch-up and there’s some interesting niche products; for example SharePoint object storage which show Dell working in partnership with ISVs to address particular pain-points but there’s nothing which shows Dell leading the way.

I think Dell need to move beyond this and think larger, they need to build a strong narrative for the future and deliver on it.

1) Dell need a strong file story, they desperately need a credible NAS product; the Exanet IP gave them a file-system but if you look at the Exanet product, that was about all it gave them.

2) Dell need an integrated management console for their products; their product range is still small enough that this is an achievable goal, if they leave it much longer, then they will stand no chance.

3) Dell need to stop comparing themselves to HP and the like; they need to build their own narrative and focus on next generation storage. This means looking at what some of the pure SSD players are doing and build a strategy around ‘What happens when SSD becomes free in the same way that spinning rust as a component is free?’; what can you do then that you cannot do now. Automated Storage Tiering is very clever but we will still care about it in 10 years or even 5 years? If all your primary storage is SSD, what changes?

4) Dell could also consider being Dell; what happens if you decide that you are not going to gouge your customer base and move away from the high margins which storage and Enterprise storage traditionally attracts?

5) Buy Brocade? Lots of smart people in Brocade and it would bring a number of innovators into the fold.

What-ever happened to Object Storage?

Jan 5th, 2012
by Martin Glassborow.
8 comments

We have heard a lot about Object Storage but really how much impact has it had on the storage market so far? EMC make lots of noise about Atmos for sure but I hear very much conflicting stories on the take-up; NetApp bought Bycast and I hear a deafening silence; HDS have HCP and seem to be doing okay in some niche markets; Dell have their DX platform and there are many smaller players.

But where is it being deployed? Niche markets like medical and legal uses but general deployment? I hear of people putting Object Storage behind NAS gateways and using it as a cheaper NAS but is that not missing the point.  If you are just using NAS to dump files as objects into an Object Store, you are not taking advantage of much of the meta-data which is the advantage of Object Storage and you continue to build systems which are file-system centric. And if you really want a cheaper NAS, there might be better ways to do it.

For Object Storage to take off, we need a suite of applications and APIs which are object-centric; we need a big education effort around Object Storage but not aimed at the storage community but at the development and data community.

Object Storage is currently being sold to the wrong people; don’t sell it to Storage Managers, we’ll manage it when there is a demand for it but we are probably not the right people to go out and educate people about it. Yes, we are interested in it but developers never listen to us anyway.

I hear Storage Managers saying ‘we’d be interested in implementing an Object Storage solution but we don’t know what we’d use it for’; this isn’t that surprising as most Storage Managers are not developers or that application-centric.

If you don’t change your approach, if you don’t educate users about the advantages, if you continue to focus on the infrastructure; then we’ll be asking this question again and again. Object Storage changes infrastructure but it is probably more akin to a middle-ware sale than an infrastructure sale.

 

Happy New Year

Jan 3rd, 2012
by Martin Glassborow.
No comments yet

Hope everyone had a nice break and is ready to get back into the swing of things; 2012 is upon us and for us living in London, we look forward to a summer of travel chaos and ever increasing levels of hyperbole. It is both the London Olympics but also the Queen’s Diamond Jubilee, so a great time to visit London and probably a great time to be living elsewhere.

Next week sees the Dell Storage Forum in London and the first #storagebeers of the year. Dell has had a year now to get their storage portfolio in order and 2012 must be the year that they begin to see their acquisitions deliver; yet, even that might not be enough and we need to see some innovation and road-maps presented. From Exanet to Compellent via Equallogic, there is enough product and I am looking forward to see how it gets woven into a strategy.

Yet Dell are not the only company who need to start weaving a strategy, arguably with the exception of EMC, this is the year when everyone needs to start drawing the weft and clothing their products with strategy and coherence.

And it is not just the vendors who need to get their strategies in order; this is very much the case for the end-user as well. Too much product and too much fluff still proliferates in many end-user organisations, this often due to a confusion between flexibility and choice.

From Cloud to Data Analytics; there has been a lot of playing with these technologies but many organisations need to move beyond this and into delivery of investment and results. As in every year, there is lots to do and as in every year, there might be too much to do. Start stripping away the fluff and delivering.

 

Dear Santa – 2011

Dec 20th, 2011
by Martin Glassborow.
3 comments

Dear Santa,

it’s that time of year again when I write to you on behalf of the storage community and beyond. 2011 promised much but delivered less than hoped; the financial crisis throughout the world has put a damper on the party and there are some gloomy faces around. But as we know, the world will always need more storage, so what do we need to deliver in 2012.

Firstly, what we don’t need is Elementary Marketing Crud from the Effluent Management Cabal; perhaps this was a last grasp at a disappearing childhood as they realise that they need to be a grown-up company.

What I would like to see is some more serious discussion about what ‘Big Data’ is and what it means both from a Business point of view but also from a social responsibility point of view. I would like to see EMC and all get behind efforts to use data for good; for example, get behind the efforts to review all drug trial data ever produced to build a proper evidence based regime for the use and prescription of drugs, especially for children who often just get treated as small adults. This is just one example of how we can use data for good.

There are so many places where ‘Big Data’ can be used beyond the simple analysis of Business activities that it is something which really could change the world. Many areas of science from Climate Research to Particle Physics generate huge amounts of data that need analysing and archiving for future analysis that we can look at this being a gift to the world.

And Santa, it can also be used to optimise your route around the world, I’m sure it is getting more complicated and in these days of increasing costs, even you must be looking at ways of being more efficient.

Flying through clouds on Christmas Night, please remember us down below who are still trying to work out what Cloud is and what it means; there are those who feel that this is not important but there are others who worry about there being no solid definition. There are also plenty of C-level IT execs who are currently loosing sleep as to what Cloud in any form means to them and their teams.

So perhaps what is needed is less spin, more clarity and leadership. More honesty from vendors and users, stop calling products and projects, Cloud; focus on delivery and benefits. A focus on deliverables would remove much of the fear around the area.

Like your warehouses at this time of year, our storage systems are full and there is an ever increasing demand for space. It does not slow down and unlike you, our storage systems never really empty.  New tools for data and storage management allowing quick and easy classification of data are a real requirement along with standards based application integration for Object storage; de-facto standards are okay and perhaps you could get some of the vendors to stop being precious about ‘Not Invented Here’.

I would like to see the price of 10GbE come down substantially but also I would like to see the rapid introduction of even faster networks; I am throwing around huge amounts of data and the faster I can do it, the better. A few years ago, I was very positive about FCoE; now I am less so, certainly within a 10 GbE network it offers very little but faster networks might make me more positive about it again.

SSDs have changed my desktop experience but I want that level of performance from all of my storage; I’ve got impatient and I want my data *NOW*. Can you ask the vendors to improve their implementation of SSDs in Enterprise Arrays and obviously drive down the cost as well? I want my data as fast as the network can supply it and even faster if possible; local caching and other techniques might help.

But most of all Santa, I would like a quiet Christmas where nothing breaks and my teams get to catch up on some rest and spend time with their families. The next two years’ roadmap for delivery is relentless and time to catch our breath may be in short supply.

Merry Christmas,

Storagebod

 

2011 – A Vendor Retrospective….

Dec 12th, 2011
by Martin Glassborow.
2 comments

So, we’re winding down to Christmas and looking forward to spending time with our families, so I guess it’s time for me to do a couple of Christmas blog entries. It’s been a funny year really, a lot has happened in the world of technology but nothing really has changed in my opinion; there’s certainly some interesting tremors and fore-shadowing though.

HP started the year in a mess and finish the year in a mess; they got themselves into a bigger mess in the middle of the year but appear to have pulled themselves from the brink of the abyss. I can still hear the pebbles bouncing of the walls of the abyss as HP scramble but I think they’ll be okay. 3Par is going to turn into a huge win for them.

EMC started the year with a Big Bang of nothing announcements and some fairly childish marketing but their ‘Big Data’ meme appears to be building up a head of steam. Isilon appears to be doing great for them and although EMC still don’t appear to understand some of the verticals that they play in now, they seem to understand that they don’t and are generally letting the Isilon guys get on with it. Yes, they’ve lost a few people but that’s always the case. Their JV with Cisco; I hear mixed reviews, I think that they are doing well in the Service Provider space but less well in the other verticals; still, they are certainly marketing well to partner organisations.

HDS still struggle around message but they seem to be getting a better selling stuff and are going aggressively after business. Much of this seems to be by ‘ripping the arse’ out of prices but a newly hungry and aggressive HDS is not such a bad thing. I still think that they are not quite sure how to sell outside of their comfort zone but some of the arrogance has gone.

IBM; Incoherent Basic Marketing. There’s a huge opportunity for IBM and yet they seem to be confused. They do have a vision and they do have technology but they do seem to struggle with the bit in the middle.  And they never seem to finish a product; so much feels half-done.

NetApp bought Engenio; a great buy but have they confused themselves? Revenues appear to be plateauing and from my anecdotal evidence, adoption of OnTap 8 is slow. I think in hindsight that some within NetApp may agree that OnTap 8 shipped too early and it was a ‘release anything’ type move; OnTap 8.1 is really OnTap 8.

Oracle ‘bought’ Pillar and still have no storage story. Larry should bite the bullet and buy NetApp; much as that might upset some of  my friends at NetApp.

I started the year with great hopes for Dell and I finish the year with some great hopes for Dell but they need to move fast with a sober HP on the horizon. HP could shut them out.

Elsewhere in the industry, pure-play SSD start-ups seem to be hot and there’s a lot of new players in that space. There’s going to be more in that space as people start to treat SSDs as a new class of storage as opposed to simply faster spinning rust. I do worry at the focus on VMware by some of these start-ups and their exposure to VMware doing something which impacts the start-up’s model and technology. Design with virtualisation in mind but ensure that you are agile enough to dodge the slings and arrows of misfortune.

One thing which has saddened me over the past eighteen months is the fall off in blog entries by some of the more notable bloggers. I know you are busy guys but is an entry every other week or so too much to ask? I miss reading some of you!! Hey, I even miss some of the heated spats in the comments.

 

← Earlier Posts Newer Posts →