Archive for the ‘PAAS’ Category

What is “The Cloud” Really?

March 4, 2010

Once upon a time I read a very good marketing paper that began with the statement: “People buy quarter-inch drill bits, but they want quarter-inch holes”.  The biggest mistake most tech companies make in marketing their products is they talk about the features of their quarter-inch drill bits, not the quality of the quarter-inch holes that can be made, or how the features of that hole are relevant or important for how the hole is going to end up being used.”  Assuming you accept this, I make the following observations about how most companies in “the cloud management space” are making it harder for their markets to understand what they do rather than easier.

Specifically, the concern I have is that “managing the cloud” or “the cloud management market” or “managing cloud computing”
is going to look markedly different depending on where you sit.  In particular, I think there are actually four cloud management markets or segments, with overlapping requirements to be sure, but still different enough that any company, vendor, or IT organization trying to “manage the cloud” should think about positioning itself in that context.  I also believe much of the confusion (or FUD) around “the cloud” and “cloud management” is because people use similar terms to mean very different things, each valid in its own right, but very, very different.

  • Segment 1 – Existing IT organizations that have on-premise services and also either have or aspires to have cloud-based services as well (whether IaaS, PaaS, SaaS, etc).  This management market will have a particular set of benefits and challenges associated with how the entity tries to integrate these IT services, and the management thereof, to make it look reasonably seamless (so they don’t simply replace one set of complex problems for a different set of complex problems).  Private/public clouds will create variations on this theme, with security and billing being the two main differences, but otherwise very similar problems.
  • Segment 2- The opposite end of this spectrum – organizations that aggressively pursue doing as much in the cloud as possible, and only doing on-premise what is either not yet available in cloud form or too business-critical to yet trust to a cloud-based solution.  I’ve spoken to a dozen CIO’s in the last two months who have set a mandate for their organizations along exactly these lines — cloud when you can, on-premise when you have to.  This is primarily an SMB-based discussion today, but it’s starting to bleed up into the enterprise space.

These first two represent more of a true user of “cloud-based” services and benefits.

  • Segment 3 – Groups that are actually hosting the cloud services used by the first two markets; the so-called “the service provider market.” It’s a real market, but tends to have a set of problems much more in common with the on-premise guys (insofar as they’re managing workloads within a well-defined IT  infrastructure — they’re still “on-premise,” just a different “premise” from the captive IT organization).  Their users come from the cloud, rather than being a captive user community.  This “where are the users coming from” tends to cause the management problems to have different priorities than the captive user version, but otherwise has more in common than not.  The one variation in this space is how high up into the stack a given organization chooses to go (IaaS, PaaS, SaaS, etc), which will also heavily influence what “management” means to them.   For example, Amazon is clearly an IaaS vendor in this space, and doesn’t know or care about applications per se.
  • Segment 4 – Also a provider market, but where all the services provided are actually located in the cloud, rather than a captive data centerConformity is representative of this type of market.  We provide a SaaS-based solution (which also happens to manage SaaS a specific problem of using SaaS applications, but that’s not a relevant distinction here) that runs entirely in the cloud, we don’t have a data center at all (except for a VPN server and a MS Domain Controller); we do everything else in the cloud (including development / builds / e-mail / calendaring / billing…. you name it).  This type of market will also have unique and real management problems, but with a very different emphasis than the first three.  It’s also still small, but rapidly growing, based on many VC discussions I’ve had in the last four months.

These last two represent more of a true provider of cloud-based services, even though they may also have “user-like” problems they need to solve.

As already noted, there is much overlap in these four “spaces,” so I don’t think one can be entirely pure in this four market segment model.  But the particular problems and their urgency will create a form of segmentation thatwill influence and govern how cloud management companies need to talk about themselves, what they do, what pain points they address, and the value of their solutions, because the value of a given solution will vary greatly depending on where in this four-market segmentation model a given customer views themselves.

So, assuming you agree at least in spirit with this segmentation, it might be helpful to start to introduce some of this nomenclature into the industry’s on-going discussion about cloud management issues.

Anyway, this is what occurred to me when I’ve tried to compare various “cloud management” offereings; it’s a bit of apples-and-oranges.  For example, CA or BMC might be expected to want to market its cloud-based offerings to companies in the first market, as on-premise is “core” and cloud is an “adjacent” space in this market (using the Baan / Zook “core/adjacency” nomenclature).

Smaller players, like UnivaUD, market its cloud-based offerings to companies in the third market, and while there’s clearly overlap and bleed-over between these two views,  they’re different enough that trying to compare them might not make sense.

As an aside, It’s also why I think “virtualization = cloud” is a horrible hoax that some vendors are foisting off on the rest of the industry.   Virtualization is an important technology, to be sure, but it’s a quarter-inch drill bit in the fullest sense of the word, and absent any discussion of what problem is being solved and why, carries almost no useful context in any final analysis of “the cloud management market.”

Thinking about “The Cloud”

February 10, 2010

Thanks, Scott, for the warm welcome to Conformity’s Blog universe.  I’ve been at Conformity for just about a month now, and I’ve been appointed (is there an opposite of disappointed?) at the excitement around the space, the quality and dedication of the team, and the interest in “our problem” (identity in the cloud) by customers and prospects.

Of course, unless you’ve been under a rock for the past, say, 10 years, you’ve no doubt heard that Cloud Computing (or On-Demand before that or ASP’s before that or Grid’s even before that) will solve everything from bad breath and world hunger to global warming and peace in our time.  While many of the developments are truly exciting, what we today call Cloud Computing should have been expected as an obvious trend from a whole collection of trends that have led up to it.

Why?  Because every advanced endeavor ultimately evolves into increasingly smaller and focused areas of specialization, where we (as individuals or business units or corporations) pay someone else to do things we’re either too busy, too inexperienced, or too lazy to do ourselves.

I suspect few of you reading this now actually grow your own vegetables.  It’s not that you can’t, mind you, since it’s not all that hard.  But farmers and grocery stores and the whole infrastructure behind the process of getting lettuce and carrots into the trunk of my car do it faster, cheaper, and better than I can (or am willing to – I do have small children, after all).

Historically, providing whatever computing services businesses large and small use in the course of their primary business activities has been difficult enough and expensive enough that these same businesses formed “IT Organizations” to provide those services for them (believing — largely correctly — that the IT group could do it faster, cheaper, and better than they could — an early and surprising enduring form of specialization).

No reason why this same process won’t happen again and again and again, with increasing segments of what has traditionally been the purview of what we now call an “on-premise” IT service being delivered by external entities that can perform more and more elements of what IT has traditionally done themselves, and with IT’s role evolving along the way.  With the introduction of a good enough transmission medium (the Internet), a good enough computing platform (LAMP stack, with or without virtualization), and sufficient consolidation, standardization, and economies of scale around certain business applications (e-mail, SFA, CRM, HR, etc), and *POOF* Cloud Computing and Cloud-based Applications are born.

The interesting news (and for companies like Conformity and our partners the good news) is that each of these forays into these areas of specialization come with their own technical and business challenges that must be solved along the way.  We, as technology professionals, get another chance to try to address long-standing questions around business process, pricing, ease-of-use, and the never-ending quest for a more efficient way to separate and distinguish between what Geoff Moore calls “core” versus “context”.

I won’t attempt to address the specifics of how we’ll be solving bad breath, world hunger, global warming, and peace in our time today (must leave something interesting to write about in future posts), but wanted to begin the dialog around what is and is not particularly new about Cloud Computing, what problems we might expect need to be solved (because they *are* different from what’s come before) and which problems are simply old wine in new bottles…

Recap: Enterprise SaaS Working Group – Identity Management in the Cloud

December 4, 2009

We had a great second meeting of the Enterprise SaaS Working Group this week, which focused on the topic of access and identity management for the cloud.  Participants in the session included Chris Bedi from VeriSign, Peter Dapkus from  Salesforce.com, Ryan Nichols from Appirio (who also provided a great summary of the event on the Appirio blog), Steve Coplan from  The 451 Group, Michael Amend from Dell, Doug Harr from Ingres and Scott Carruth from Initiate Systems.   Our initial discussion focused on the unique management challenges created by SaaS and cloud applications due to the the identity silos they create in the enterprise as shown below.

Cloud identity in the enterprise

The ensuing roundtable discussion focused on the impact these issues are having in the enterprises, with a particular focus on the following topics:

  • Speed bump or show stopper – on the question of whether access and identity management issues were a going to be a ‘speed bump’ or ‘show stopper’ for SaaS adoption in the enterprise, the answer really revolved around timing and depth of penetration.  While today it is more of a speed bump for initial adoption in the enterprise (or else we wouldn’t be seeing enterprise deals today), the issues become more problematic when considering what it will take for SaaS and cloud applications to become a ‘mainstream’ technology. Taken from that perspective, there was agreement that identity issues around access, authentication and authorization created by SaaS identity ‘silos’ were going to soon become major, and that they need to be reconciled and addressed.  
  • The directory redefined – one of the questions we posed around the future of the corporate directory, and whether enterprises would ever permit it to live in the cloud.  Chris Bedi of VeriSign made the great point that the more relevant and important question is around what a directory really becomes in a cloud-centric environment – where it ends up residing will be a function of how that question is answered.
  • Federated identity – related to the directory point, the group generally also agreed that in a cloud-centric (or even hybrid SaaS/on-prem environment) that there was unlikely to be a monolithic directory or source of identity related data, and that SaaS applications, HR systems and directories (on-prem and cloud) would also likely each contain ‘versions of the truth’ that will need to be synchronized and federated.  Ryan Nichols provided a very interesting example of how Appirio themselves have built a cloud-centric organization with Salesforce.com and Google both providing separate but complementary directory and identity data.
  • Identity done right – Doug Harr made the excellent point that current cloud identity challenges actually offer an opportunity for SMB and midsize enterprises who haven’t been able to invest in identity and systems management technologies to date to ‘get it right’.   IAAS and cloud-based identity management services will likely make these capabilities cost-effective for these target markets for the first time, enabling these organizations to effectively ‘white sheet’ their identity management approaches for both cloud and on-premise applications.

The full recording of the webinar is available and can be access by clicking here.  Please drop us an email as eswg@conformity-inc.com to be added to our mailing list, and to be notified of future Enterprise SaaS Working Group news and events.

Recap: The Enterprise SaaS Working Group

October 1, 2009

It’s been an exciting few days here at Conformity after our recent GA announcement and the kickoff of the Enterprise SaaS Working Group yesterday.  We had a very lively, engaging debate on the key issues the group believes need to be addressed for SaaS and cloud applications to become ‘mainstream’ technologies in the enterprises.  The group featured a diverse set of executive perspectives from cloud vendors, thought leaders and practitioners, and included:

A quick highlight of some of the discussion yesterday:

  • PaaS/SaaS – which model ‘wins’ in the enterprise? While opinions differed, a common sentiment shared by the panel was that there’s not going to be ‘right answer’ for all organizations.  Depending on the industry vertical, business process or IT management model PaaS or SaaS could be the ‘right answer’, and in many situations organizations could have PaaS and SaaS offerings sitting side by side.   
  • Private clouds – part of the answer or indicative of SaaS market immaturity? As with the PaaS/SaaS discussion a common theme was ‘it depends’.  The core advantage to SaaS and cloud delivery models is the ability to share resources – what part of the stack organizations decide they’d like to share will likely be driven primarily by security concerns and issues.  A likely scenario, as with PaaS/SaaS, is that different models will likely be adopted by different types of organizations depending on security and operational requirements.
  • Enterprise SaaS adoption – when does it overtake on-premise? Two different perspectives were discussed around when SaaS will overtake on-premise apps in the enterprise.   A common belief of the group was that SaaS is winning in a majority of new deals in the enterprise today, with the perspective shared that 50-75% of enterprises would ‘flip the switch’ on cloud in some manner by approximately 2012.  Peter Coffee of Salesforce also shared his belief that total installed base for SaaS would outnumber on-premise apps by 2020, though there would also likely be 1-2% of the market that would be ‘holdouts’.
  • Any applications that SaaS/cloud won’t be able to penetrate? If architected and deployed correctly, there are no perceived areas in which SaaS and cloud application models could not be leveraged with Peter Coffee of Salesforce , Tom Fisher of SuccessFactors and Ryan Nichols of Appirio all providing compelling examples of large scale, transaction intensive customer deployments.

The full recording of the webinar is available and can be access by clicking here.  Also, Ryan Nichols at Appirio had a great post on their perspective on our discussion topics here.

Please drop us an email as eswg@conformity-inc.com to be added to our mailing list, and to be notified of future Enterprise SaaS Working Group news and events.

Enterprise-Class SaaS Provisioning

June 3, 2009

As those of us at Conformity engage enterprise IT teams, we continue to explore the gap between existing provisioning options and SaaS deployments.  Enterprise customers are caught between the promise of cloud and SaaS solutions and the impact of this adoption on their already stretched teams and processes.   In the Conformity white paper, Enterprise-Class SaaS Provisioning, we describe the management challenges organizations face in adopting SaaS applications, and explain why IT groups struggle to utilize existing options for federating on-demand environments.

So, what information can we take away from the enterprise SaaS customers?  As pointed out in our other discussion threads, SaaS is not easily tamed by existing solutions.  We find that the cloud deployment model exposes the following shortcomings of existing alternatives:

  • Disconnected Environments: The most obvious challenge is the separation of multiple SaaS applications and the management solutions.  This disconnect fragments the core IT capabilities, creating unique cloud-based silos of user identity, business policy, and administrative rights.
  • Unexpected Deployment Complexity: IT teams can easily underestimate the impact of adopting SaaS as a solution platform.  Detailed SaaS configurations, coordination between applications, and evolving licensing models can exceed IT expectations, especially when the deployments were independently cultivated in the lines of business.
  • Lack of Deployed Standards: Customers are discovering the industry standards for management and provisioning are not aligned with the aggressive SaaS expansion.  Many advertised standards such as SAML or XACML are focused on alternative use cases and designed for either an on-premise or cloud model, limiting their real adoption by SaaS ISVs.

These challenges have curtailed enterprise efforts to utilize current deployed technologies, and in turn have impacted SaaS rollouts.  IT teams continue to evaluate complementary but incomplete options including enterprise software vendors, cloud-based identity solutions, and unique SaaS ISVs themselves.  This discovery process has provided the benefit of allowing the enterprise teams to better understand the market challenges and applicability of existing solutions.

Working with these IT teams, we have defined a common set of issues for provisioning and management and select criteria for a new approach to federating on-demand environments.  Any solution must provision users to a fully functional state across the user life cycle, a distinct challenge with many SaaS and cloud implementations.   This provisioning must align with existing IT and business processes, leverage line of business expertise, and meet the organizations compliance, security, and data visibility needs.  And deployments must be flexible enough to align with and possibly impact developing standards such as SPML or federation options like Microsoft Geneva while supplying value prior to standards adoption.  In short, these attributes define a new breed of management platform that is designed for the SaaS and cloud-based environments.

For more information, read the Conformity white paper that outlines our findings.  And please feel free to reply and continue the discussion.

The Open Cloud Manifesto – there’s work to be done…

April 3, 2009

Without commenting on the motives of the players involved in  The Open Cloud Manifesto published this week, we do have to agree with one of its core tenets – that to drive further adoption and acceptance SaaS and cloud providers must work together to improve overall governance and management of offerings.  At a fundamental  level SaaS and cloud-based applications ‘break’ the models and approaches organizations have implemented for managing users, identity and applications in a primarily on-premise world (particularly in large enterprises).   SaaS and cloud based applications nearly all have their own unique, individual approaches for managing users, profiles and permissions, and do not easily integrate into existing management solutions and directory services.  To date the only standards that emerged in the broad area of  ‘management’ address access and authentication issues (SAML, OpenID etc), and none of these have even gained significant traction with SaaS ISVs yet.   When looking at the new broad, cross-vendor governance and management issues created in a multi-SaaS environment, access and authentication is only the tip of the iceberg, which is why we also believe there’s significant work to be done here…

A big week for SaaS

September 21, 2007

Quite a big week for SaaS, with both Salesforce.com and SAP making major platform and product announcements…

At their annual user conference Dreamforce, Salesforce.com introduced their new Force.com application development platform. The new ‘platform-as-a-service’ or PAAS offering (yes, another new SaaS-related acronym) extends the AppExchange environment, providing users a platform and toolset that enables them to leverage UI, logic, database and integration ‘as-a-service’ components. An early peek also was provided at VisualForce (currently only available as a developer preview), which will enable the development of custom UIs on the Salesforce.com platform. On the application side, Salesforce.com appears to be continuing its strong focus on the CRM market.

SAP also unveilved (finally) Business ByDesign, its highly anticipated on-demand mid-market ERP solution which was formerly know as A1S. With standard pricing at $149 per user per month, the jury is out around the cannibalization potential of BBD on SAP’s ‘legacy’ business, and the resulting impact that might have on SAP’s excitement around and commitment to the SaaS model.

The net net? The addition of another major player to the SaaS market and the introduction of a new platform and tools to enable the rapid development of new SaaS applications – more choices and options for organizations seeking to leverage the SaaS model, and more to eventually manage as well…