lustratus REPAMA

Posts Tagged ‘taxonomy’

I was reading Robin Bloor’s blog entry from earlier in the year recently where he makes some interesting points. But I’m not sure that I agree with all of them…

Robin attempts to identify some of the defining characteristics of cloud computing but I get the feeling that he is starting from a slightly cynical stand point. And whilst I agree that vendor and service provider marketing tactics have created a great deal of hype, I feel others must also share some of the blame for the confusion in the space.

Anyway, his 10 Defining Points for Cloud Computing blog entry is here. I tried to post a comment but as it appears that it’s still awaiting moderation, I thought I’d reproduce it below.

An interesting read as ever Robin but I can’t help thinking that you’ve defined Cloud Computing more so by the things it is not, than what it is. But perhaps that is the place to start when defining something as nebulous as Cloud Computing.

Completely agree that cloud computing is a confused set of definitions and misunderstandings but whilst I acknowledge that the marketing tactics of vendors and providers should take some of the blame – so should the market analysts. The global IT analysts have resisted the temptation to coalesce around a series of definitions and market categories for reasons of vested interest. De facto market category definitions and market segmentations are still up for grabs. So for analysts it’s currently about land grab – attempting to be the firm that defines the broad Cloud Computing categories and drivers. Their attempts to define and steer the market on their terms simply further serves to confuse prospective Cloud Computing users.

I also agree that standards are currently few and far between. The upshot of this is that market analysts and vendors can afford to go in different directions without being restrained by common understanding and artificial technical limitations. Again the impact is on the potential Cloud Computing user who is left trying to find their way in a world that is not easy to define, classify and compare.

Finally the ease at which technology vendors can enter into a hosting agreement with someone and then rename a product line and voila – they become a Cloud Computing vendor, has made the Cloud Computing market incredibly crowded, incredibly quickly. Attrition and customer cynicism should account for the weakest vendors here leaving the candidates for success in the market categories in which they compete.

For what it’s worth I’ve tried to bring some sense to the categorisation of the various market categories and cloud computing vendors/providers in the Lustratus REPAMA market landscape, taxonomy and segmentation model. Instead of ploughing my own furrow, I stood on the shoulders of giants and worked from some of the best definitions I had found in the blogosphere and beyond.

I wondered if your readership my find this useful. This is discussed here.

http://www.lustratusrepama.com/go-to-market/a-market-landscape-for-cloud-computing/

My colleague Steve Craggs has also attempted to define cloud computing in lay-terms based on the categories above.Steve’s piece is discussed here.

http://www.lustratusrepama.com/marketing/cloud-computing-explained-without-the-hype/

Danny Goodall

My Lustratus Research colleague Steve Craggs has taken a step back from the hype surrounding Cloud Computing and has defined it for the layman.

Steve’s premise was that the Cloud Computing hype had reached fever pitch and that some of the claimed benefits had become pretty difficult to believe. He wanted to strip it down to the basics, define it and look dispassionately at the likely benefits and the trade-offs involved. Steve has also put some of the vendors and service providers in the space into the right context based on the Lustratus REPAMA Market Landscape / Taxonomy / Segmentation model.

The document is embedded below and can be downloaded from Scridb or the Lustratus web store.

Danny Goodall

Steve and I had an interesting and refreshingly different call this morning with GigaSpaces.

They had reached out to us to make sure we knew all about them so that they could be fully represented in each of the categories in our cloud computing market landscape / segmentation / taxonomy where they have a solution.

And after the call I can see that instead of just being in

• Cloud Software / Compute

They also have a valid claim to

• Cloud Software / Data

and may be even

• Cloud Software / Cloud Management ( /Application Services Management )

This raised a bit of a dilemma for me. I certainly want to represent vendors and providers accurately in the segmentation model but I want to avoid vendors appearing in lots of different segments just because they believe that some esoteric feature or other qualifies them.

Instead I want the segmentation model to reflect where the vendor/provider specifically and actively addresses a market need with a specific capability and value proposition. So if I hear from a vendor/provider “yes we can do that too”, I’m reluctant to simply add them to a segment. After all the motivation behind producing the segmentation model was to remove some of the confusion present in the Cloud Computing not to perpetuate it.

So I think we’ll need to reflect on what we heard this morning from GigaSpaces and think how best to represent those vendors in our market segmentation document that have a single product with broad capabilities. Having said that I think there is enough about the XAP proposition that means I will be adding it to at least one more segment of the market landscape.

Anyway, the refreshing part of the call was that GigaSpaces’ marketing seems to focus on what they can do for organisations rather than simply placing the “Cloud Computing” term before, after and in the middle of their product name everywhere it appears. In fact you have to go digging on their web site to find references to the cloud-enablement features of their product line. I’m not sure whether they have taken this approach consciously or if it is that they’re not sure how best to position their offering in the cloud market. The risk I guess is that as they are not positioned specifically as a cloud computing vendor, they may not be placed on clients’ long lists. But it certainly differentiates them and different is usually good unless you’re selling 8-fingered gloves.

These calls with vendors are useful for many reasons but mainly to test some of my assumptions. As with many of the products and companies I have looked at in compiling the segmentation model, I had “assumed” that I knew what the product was and what it did. It turns out there are many more strings to XAP’s bow than I had at first realised. I knew GigaSpaces as a purveyor of extremely scalable application servers but that, it turns out, is only half the picture.

One interesting feature of GigaSpaces’ XAP product is its application services management layer (my term not theirs – they use application management services!). This layer understands service level commitments for the infrastructure as well as the applications that are deployed to it. It’s common practice for the infrastructure to “understand” the service level that is expected from the infrastructure itself – usually measured in CPU percentages, data volumes or some such. However it’s not so common for the infrastructure to understand the commitment that the application developer has made about the service level that the application will deliver. These application layer service levels are usually described in some business metric and to then have the infrastructure react to automatically provision more infrastructure to meet the business’ requirements to ensure that application SLA commitments are met is a certainly an interesting claim. One that we will dig into a little further over time. GigaSpaces’ Jim Liddle explains a little more here – whist also picking a fight with TIBCO over whether its claim that “self-aware elasticity” is something unique to Silver.

Interestingly, it appears that with a number of GigaSpaces recent Cloud Computing customer wins (they claim to have 75+ “cloud” customers) they have in fact used this application management services layer to control and manage Amazon’s EC2 infrastructure. This ability to manage application services outside of its own compute infrastructure goes some way to answering my rhetorical question of whether any vendors were focussed on providing a pure play layer for application services management. It appears that there are.

So I’ll make some changes to the segmentation model in light of our conversation this morning and I’ll also keep an eye on how GigaSpaces’ cloud proposition develops – specifically following their Platform as a Service partnership with ServePath’s Go-Grid announced earlier today.

Danny Goodall

I’ve completed the first draft of the cloud computing segmentation model upon which we will build our REPAMA studies.

As I’ve mentioned before along my journey to arrive at this model, I’ve found the cloud computing market to have quickly become crowded and confused. This is largely due to the ease at which “traditional” vendors have re-repositioned themselves to catch the cloud computing wave.

The other issue of course is that over time cloud computing will cease to be a new paradigm and will quickly become the way consumers and businesses avail themselves of computing services. So what I’m seeing here is a market in transition where just about every category in traditional software sales will have an offer in the cloud computing space until on-demand models becomes “the norm”.

So I guess it’s really not that surprising to see so many vendors present in the space. But at the same time it is very confusing for legitimate prospects to cut their way through the mass of terminology to then examine vendors and service providers who appear to have broadly identical capabilities and value propositions.  How do they decide the best way to take their first steps into cloud computing? It’ll be interesting to see what our REPAMA studies say about how each of the vendors/service providers’ takes their products to market.

Anyway, I’ve uploaded a set of slides to slideshare.net which I think is probably the best way to make the material available but if anyone wants a copy of the slides please let me know. The slides are embedded below.

As I’ve said before, this segmentation model will undoubtedly develop and change over time as I look in more detail at the marketing efforts of the various vendors involved. The definitions for each of the functional areas are a little woolly right now. But at least I now have a structure that allows me to decide which segments and vendors/service providers I will include in our studies moving forward.

I’d like to once again acknowledge the significant role that Brad Buck, Peter Laird and Christofer Hoff played in helping to form the ideas on market segmentation and the role NIST has played in crystallising definitions on cloud computing and software/platform/infrastructure as a service.

Danny Goodall

Products and vendors included in the segmentation model are shown below. If you represent a vendor below and I haven’t represented your organisation correctly, or if you represent a vendor that isn’t included but should be, please contact me and let me know a little bit about your company and your proposition and where you feel you fit in the segmentation model.

10Gen MongoDB, 3Tera App Logic, Aconex, Advologix, Altor Networks, Amazon EBS, Amazon EC2, Amazon S3, Amazon SimpleDB, Amazon SQS, Amitive, Apache CouchDB, Apache HBase, Appian Anywhere, Appistry, AppJet, AppNexus, AppZero, Aptana, Aria Systems, Aster DB, Beam4d, Beowulf, Blink Logic, Boomi, Box.net, Bungee Labs Connect, Caspio, Cassandra, Cast Iron, Clickability, Cloud42, Cloud9 Analytics, CloudFoundry, CloudStatus, ClusterSeven, CohesiveFT, CohesiveFT VPN Cubed, ColdLight Neuron, Collabnet, Concur, CrownPoint, CTERA, CTERA Portal, DataSynapse, Desktoptwo, DirectLaw, DocLanding, DropBox, Dynamsoft, Dynect, Elastichosts, Elastra, EMC Atmos, Engine Yard, Enomaly Enomalism, enStratus, Etelos, Eucalyptus, eVapt, FathomDB, Fios, Flexiscale, Force.com, Gemstore Gemfire, Gigaspaces, Globus Toolkit, gnip, Google App Engine, Google Apps, Google BigTable, GridLayer, Hadoop, Hosting.com CloudNine, HubSpan, Hyperic, Hypertable, IBM Lotus Live, iCIMS, InfoBright, Informatica iTRICITY, Joyent Accelerators, JungleDisk, K2 Analytics, Kaavo, Knowledge TreeLive, LayeredTech, LiveOps, LoadStorm, LogiXML, LongJump, LucidEra, memcached, Mercury, mezeo software, Microsoft BizTalk Services, Microsoft SDS, Mosso Cloud Files, Mosso Cloud Servers, Mosso Cloud Sites, Mozy, MS Azure Services Platform, MSDynamics, MuleSource Mule OnDemand, NetDocuments, NetSuite, NewRelic, Ning, Nirvanix, Oco, Open.ControlTier, OpenCloud, opencrowd, OpenNebula, OpenQRM, OpenRSM, OpSource, OpSource Connect, Oracle Coherence, Oracle On Demand, Panaroma, Parallels, ParaScale, Parature, PingIdentity, PivotLink, Platform, Qrimp, Quantivo, Questys, rackspacecloud, Redi2, Reductive Labs Puppet, Responsys, Rightnow, RightScale, Rollbase, rPath, Salesforce.com, Scalr, Sertifi, Serve Path GoGrid, SkyTap, SnapLogic, SnapLogic SaaS Solution Packs, SOASTA, SpringCM, Sterna, StreetSmarts, Success Metrics, Sun Grid Engine, Symplified, Syncplicity, Taleo, TerraCotta, Terremark, TIBCO Silver, Tokyo Cabinet, Trigence, Vertica, VMWare vSphere, Vordel, Workday, Workxpress, Xactly, Xero, Xeround, Xythos, Ylastic, Zembly, Zmanda, Zmanda Cloud Backup, Zoho, Zuora, Mezeo Software, Workxpress, Trigence, AppZero, Platform, OneNetwork, SpringSource, Vaultscape

Continuing my quest to segment the cloud computing market, I’m now looking at the role a channel might play in cloud computing…

…and I’m struggling a little to map the traditional channel role onto cloud. But here are my current thoughts.

There are some obvious areas in the cloud taxonomy/segmentation that look like a good old fashioned software sales model. So first let’s start with my draft / work in progress taxonomy/segmentation to help anchor the discussions in something solid. BTW draft/work in progress means that it will change.

Cloud Software

So the Cloud Software segment looks like a traditional software business. Using Brad Buck’s definition for this segment:

Cloud Software is off-the-shelf software that can be used to create an internal cloud or in some cases can be used to customise infrastructure services to mould a customer cloud solution.

Although Brad implies this, I’d extend that even further to say that cloud software will be the basis upon which infrastructure service providers build, host, manage, secure, monitor, etc. their own infrastructure services. So as this is real software being sold to real organisations the channel model I’ve come up with looks familar.

Software is sold to an organisation either directly or through some form of reseller/distribution channel. The end user IT organisation making the purchase either creates a cloud on their own premises for internal use (internal cloud) or a cloud service provider uses the cloud software to create a cloud service which is then made available to others.

Software Services

Looking at the software services segment, the definition here is

Software services are applications or components that can be used as an end application or used as part of a custom solution

So we have packaged applications or application components of varying degrees of complexity made available to application users directly by the software services provider. Application builders will incorporate some of these software services into their own applications which are tailored and then made available to application users – thus forming one channel to market.

The draft channel model I see there is as shown below:

Here we see that the application user is the ultimate consumer of the software service and that they are likely to consume that directly via the software service provider or via a third party application builder that modifies and extends the software service.

Platform Services

Broadly similar to the platform as a service definition, this segment is defined here specifically as:

Platform services offer a ready built infrastructure and application framework than can be used for building and running applications.

Platform services provide the frameworks in which applications can be built, tested and run. The application user is ultimately the end consumer of the service but some form of application builder will have used the platform service to make the application available to the user as shown below:

There is an argument here that the strict consumer of the platform service is the application builder who then makes the application available to the application user as a software service model. I need to think about this a little more before I cast this section in stone.

Infrastructure Services

Finally the infrastructure services segment is defined as:

Infrastructure services provide building blocks that can be moulded to run different application servers, packaged applications, grids, etc., which can be used to host applications.

These are basic data centre-like hardware, software and network elements that are provided as building blocks upon which software services and platform services can be deployed. In addition end user IT organisations also procure infrastructure services directly and then deploy their own software on to it, treating it as a virtual part of their own infrastructure.

This means that the channel model for infrastructure services looks like this:

Implied Channel Hierarchy

There is an implied channel hierarchy as shown in the diagram below but this is not always strictly followed. The diagram suggests that each level of service provider engages the services of the higher level service provider.

Whilst this certainly will be the deployment model for many cloud users it will not always be the the case. For example whilst it might make academic sense to assume that software services are built and made available using platform services which are then deployed in infrastructure services, in reality services at each level can be provisioned and delivered autonomously without the need for a relationship with providers at any of the other levels.

Conclusion

Well I can’t say that I’ve reached any sort of conclusion yet but I think getting these drafts down on paper has helped me think about the way that the IT industry’s traditional channel models will change over the coming years. I think it’s clear that we will either have less “middle-men” in our sales models or that the roles of these middle-men will change significantly as cloud computing becomes more prevalent.

OK, so back to the segmentation exercise. Next I need to decide on the candidate sub-segments within the major segments explored above. And then I need to decide on which vendors will appear in each of those sub-segments. Finally I need to decide which segments, sub-segments and vendors/providers will be the focus of our cloud computing REPAMA research. I’ll keep you posted on my progress.

The  DRAFT Channel Models for Cloud Computing presentation from which the diagrams above have been extracted is shown below.

I’d like to once again acknowledge the significant role that Brad Buck, Peter Laird and Christofer Hoff played in helping to form the ideas on market segmentation.

Danny Goodall

I’m continuing the REPAMA Segment Analysis Study into the Cloud Computing market attempting to arrive at a solid market segmentation and two things have become very clear.

Firstly, every vendor with a remotely related proposition appears to have added the word “cloud” to their product name, presumably in an attempt to bask in the reflected glory that cloud computing provides, perhaps in an effort to appease their investors. This means that there are a large number of vendors claiming to be part of specific segments that may or may not have legitimate claims. This makes the process I’m going through confusing and messy. And if I, as a marketing analyst am having problems, I wonder what sort of success a legitimate prospect would have in finding the product/service they need.

Secondly, the market is still in flux and as such nailing a clear segmentation down is, for some time, going to be like herding cats. This is a classic early market symptom where the market is currently led by vendors’ own ideas about what is needed and what is possible. Once the market starts to form properly and vendors align behind what their prospects and customers are actually buying, then the segmentation will become clearer.

But I have to start somewhere and the good news is that I’ve been corresponding with both Peter Laird of Tendril Inc. and Brad Buck of OpenCrowd and they both have given their blessing to this project and have offered to help if they can. Most importantly they are both happy for me to use their work on a cloud computing taxonomy/model, here and here, as a starting point for my cloud market segmentation.

One thing that I quickly realised was needed for my purposes was some form of professional services (human skills) offer. Plenty of large consultancies, smaller integrators and boutique IT shops offer consultancy services around cloud computing. Both Peter and Brad have rightly focussed on the categorisation of product capability and customer need in fleshing our “their” taxonomy. But to fully understand what propositions are being put together to service the cloud computing “need”, I must include cloud computing professional services.

So far I’ve identified the following list of services under the category of professional services. I suspect that it will grow some and be rationalised some before I finish.

• Strategy, planning and design
• Migration and implementation services
• Testing
• Security
• Cloud application design/porting
• Support services and training

I suspect that cloud application design/porting will be a relatively niche category for a while and that testing, security and support services may all become amalgamated as I carry on my research.

Another major category of cloud computing offer that has also presented itself is that of “Channel”. A number of organisations are white labelling, OEM’ing, reselling or otherwise fronting cloud vendors/service providers’ offerings. One concern that I have is that this sort of partnering is very common in hot early markets where one vendor with no cloud proposition hastily signs a partner agreement to simply tick a box. It does’t mean that either company will do any real business.  I’ll do some digging and and if it looks like a real enough category I’ll add it and blog on it soon.

Danny Goodall.

In putting together our REPAMA analysis into the go-to-market strategies of the vendors in the cloud computing space, we first must arrive at an agreed segmentation of the market. This blog documents that process.

OK so as I mentioned here, I am going to carry out a series of REPAMA Segment Analysis Studies into the cloud computing market. The desired end product is a series of reverse-engineered go-to-market strategies for a set of vendors in each of the categories within the cloud computing market. But first I need to decide on the segmentation of the various technical offers in the cloud computing space.  After that I need to decide which vendors fit into each of the different proposition segments. This series of blogs will capture my journey through this analysis.

Where to start? Well I thought that my colleagues Steve Craggs and Ronan Bradley at Lustratus Research (the market analyst part of Lustratus) would be able to give me a definitive answer on the segmentation of the cloud computing market but as Steve told me, right now there isn’t a universally agreed way of describing the various categories of propositions under the cloud computing banner. Whilst he feels their are some obvious high-level classifications, under there things are a bit grey.

The problem is two-fold. Firstly, it appears that as the cloud computing market is currently at an early stage (albeit using what now are some pretty mature technologies); the propositions have grown, merged, change direction and are only now starting to find live customers amongst enterprises; so categorisation has proved difficult.  Secondly, the analysts involved in this space each has a vested interest in classifying the market on their terms and importantly in a different way to their analyst competitors – a game they will play until the market matures.

So I decided that I would capture the opinion of the great and the good of analysts and vendors alike and come up with our own classification. And in doing this a number of existing works stood out.

1. Peter Lairds – On-Demand Blog Classification

Peter Laird’s “Laird On-Demand blog“, and in particular this post, documents his efforts to arrive at a market taxonomy for Cloud Computing together with a nascent list of the vendors that fit into each category. As you can see from the diagram on the right, Peter is a fellow fan of mind mapping (it’s how Lustratus collects, collates and generally makes sense of the raw data behind the REPAMA studies), his classification has a simplistic appeal that I like.

At the highest level Peter divides the cloud market into

• Infrastructure
• Platform
• Services
• Applications

I feel that this a good way of looking at the cloud market and one that works well for my purposes. I also like the way Peter has classified private and public cloud seperately.  From the perspective of the REPAMA analysis I will be conducting, this is an important distinction between the various vendors’ propositions. (This despite the fact that the same software and vendors may appear in both categories).

2. OpenCrowd Taxonomy

Another interesting approach is provided by OpenCrowd.  I don’t know OpenCrowd but they appear to be a RIA vendor but also deep thinkers who really understand the major trends in the markets in which they compete. They’ve looked at cloud computing and have come up with a very thorough classification as shown in the diagram on the right.

They see the market also categorised into 4 high-level categories:

• Infrastructure Services
• Cloud Software
• Platform Services
• Software Services

OpenCrowd has also gone to considerable effort to sub-categorise these high-level categories and also to identify specific vendors associated with those sub-categories. Again I like this approach for its thoroughness although having looked at some of the vendors in each of the categories, its clear that they have been placed there/asked to be placed there simply because they want to bask in the reflected glory of the cloud computing market, as opposed to having a dedicated or specific cloud focus.

3. Christofer Hoff’s Cloud Model

Finally, a more technical classification comes from Christofer Hoff in his highly entertaining and incredibly well thought out security-focused blog – Rational Survivability. In this entry Hoff publishes his architectural model which appears to be largely aimed at understanding the interaction, dependencies and relationships between the architectural components in a cloud architecture from a security perspective.

It’s useful to have an architectural model to work with as this helps to validate that the vendor-led offers into the cloud market as described above, have some basis in fact. Whilst it won’t figure in my market-led classifications I have added it here for completeness.

So I’m still in the process of collecting my ideas and I’d like to take bit of 1. and 2. above and merge them into something that better reflects the vendors’ perspectives of how they take their products to market. I also need to first ensure that both Peter and OpenCrowd are comfortable with me doing that. The model I arrive at will necessarily be far simplified too as for the purposes of my vendor to vendor comparisons, I need to ensure that I group vendors together that compete even if they don’t intellectually appear to fit into the same market categorisation.  Watch this space.

Danny Goodall.

• Infrastructure
  Public
  Private
  Infrastructure
  Platform
  Biz Users Platforms
  Dev Platforms
  Services
  Storage
  Integrate
  Enablers
  Applications