Anyone that is still vociferously doubting and denying the future of cloud computing and its near-mainstream nature will sound as strange and nutty as some US Senate hopefuls that still proudly deny evolution and climate change (while admitting to “dabbling with witchcraft” in the not-too-distant past). In fact, can anyone name a renowned enterprise resource planning (ERP) vendor that has not yet at least announced its cloud computing plans and strategy (if not already delivered actual cloud products)?
During the Grape Escape 2010 event this past summer, the common theme in all four featured vendors’ announcements was getting the “cloud religion.” I am still amazed to see how some of these vendors’ mantras have transformed from “Our customers do not ask for it!” to “We are in the cloud too!” in just a couple of years.
Indeed, it has been interesting and satisfying for me to watch the increased traction and shifts in company size, application footprint, and vertical sectors that are increasingly looking to cloud and software as a service (SaaS) technologies. Again, all that mindset change has taken place in a matter of years, from the time when the concept was both new and prescient.
There are many definitions of cloud computing, but let me try to paraphrase Gartner and Saugatuck Technologies, given their vocal work in this area. Namely, cloud computing is a style of computing whereby scalable and elastic IT capabilities are provided as an on-demand service to internal and external customers using Internet Protocol Suite technologies. Service-oriented architecture (SOA) is not necessarily required for cloud computing, but SOA-based enterprise applications make adoption of cloud computing much easier.
In cloud environments, IT resources (whether hardware, software, or both) are shared and consumption can be metered by use if required (although, more often a monthly subscription fee per user covers “all you can eat” consumption). Although cloud computing is based on Internet protocols, it is not necessarily based on the Internet per se, as much cloud computing takes place in virtual private networks (VPN) and intranets that are independent of the public Internet.
Single-tenancy vs. Multi-tenancy and Private vs. Public Cloud Debates Continue
Needless to say, not all clouds are the same, and I don’t refer here to the meteorological cirrus, stratocumulus, or cumulonimbus types. In the world of enterprise software, one can, e.g., differentiate between private and public clouds. Private clouds are perceived as more secure, since they lie behind the corporate firewall and utilize the same security tools as the company’s existing (supposedly secure) intranet. Many companies tend to look at private clouds first, as they seem safer and can save them some money quickly, especially if the company has not taken the virtualization route yet.
Public cloud computing is also safe for certain activities, such as evaluating new software. For many enterprises, public cloud computing is also appropriate for application development (and testing) and simulations (and modeling) that require extreme scalability and elasticity. Moreover, many enterprises are putting live production data, transactions, and workloads into the public cloud, as cloud computing and its vendors are maturing rapidly. As best examples, one can think of salesforce.com, SuccessFactors, ADP, and NetSuite’s sizeable customer communities in the public cloud.
My colleague Jorge Garcia’s recent tutorial article explains well how to approach cloud computing from a down-to-earth, practical point of view. Furthermore, Jorge’s more recent blog post also attempts to draw a demarcation line between traditional application service providers (ASPs) and contemporary SaaS providers.
In addition, my recent blog series on Consona’s acquisition of Compiere offers and in-depth “pro et contra” analysis between a private (single)-tenant cloud approach vs. a multi-tenant cloud approach. In a nutshell, the fundamental difference comes down to the flexibility and malleability of the cloud software by the customer (in case of private single-tenant cloud setups) vs. the cost and economy of the scale of multi-tenant public environments (but often at a “one size fits all” expense).
In his recent, somewhat provocative blog post, Josh Greenbaum argues that multi-tenancy is more beneficial to a vendor per se than to a customer, which has prompted some multi-tenancy and public cloud proponents to respectfully disagree. Somewhat ironically, muti-tenancy can be nuanced on its own, given that some vendors tout their multi-tenant SaaS solutions being delivered as a global single-instance, contrary to many multi-tenant multi-instance offerings.
In other words, salesforce.com and Plex Systems, e.g., deliver their software from a single place, which ensures that each customer on earth is on the exact same version of software at all times. That is not necessarily the case with other vendors, e.g., SAP Business ByDesign is multi-tenant but is (or will be) delivered by partners and from the vendor’s multiple regional sites. In that case one cannot be sure that every customer is on the exact same version of the software at all times. But that particular debate will come in some other article.
PaaS the Platform and Infrastructure, Please
To confuse end-users even more, pundits have created a few more acronyms based on which layer of the technology stack is provided in the cloud. To that end, one can think of infrastructure as a service (IaaS), platform as a service (PaaS), and SaaS. Simply speaking, PaaS can also be called “middleware as a service” as it is the middle layer in the big-picture architecture of an entire technology stack in cloud, with the infrastructure layer (IaaS) below it and the applications layer (SaaS) above it.
In fact, all the middleware tools used in traditional on-premises software deployments are needed in the cloud as well. There are thus several categories of middleware available as a service, starting with enterprise applications integration (EAI) as a service. For example, some IT organizations have been doing advanced business-to-business (B2B) integration and messaging projects via IaaS.
For clarity, I will refer to PaaS here only to denote an application development platform in the cloud, such as Force.com by salesforce.com, Windows Azure Platform by Microsoft, or Google App Engine. The application platform, whether in the cloud or on-premise, is where the programming model, business logic, and basic architecture of applications are determined. Currently, in terms of internal IT departments, PaaS is most appealing to large projects on a fast-track, small projects with limited budgets, and pilot projects.
Many Roads to Cloud ERP
There are no established PaaS leaders yet, but salesforce.com, Google, Microsoft, NetSuite, IBM, Oracle, Progress Software and a dozen smaller players (some of which are open source) such as LongJump, Ruby on Rails, Python, and Engine Yard are making strategic investments in their PaaS offerings. Over time, I would not be surprised to see SAP join the PaaS fray, as well as a number of emerging open source alternatives that could rival the approaches taken by their proprietary counterparts.
Most independent software vendors (ISVs) already offer or plan to offer some or all of their applications as a service (on-demand software). To do so, they have to find an appropriate way to convert their applications to the cloud model.
Some vendors will simply use traditional hosting services or IaaS (most often via Amazon Web Services) and make no architectural changes to their applications. This “same old SaaS (soSaaS)” stopgap approach is not sustainable in the long run, since it does not have the same economies of scale, agility, extensibility, and its implementation-centric architecture makes it incapable of delivering equivalent business value.
On the other hand, other vendors will try to rewrite their applications in a multi-tenancy manner themselves. For example, SAP has lately made significant investments across the entire PaaS and IaaS stack, as the SAP Business ByDesign product seems to be ready for prime time. But SAP would also be a textbook example of why this approach can work, but it is expensive, time-consuming, and gut-wrenching, and creates long-term burdens. Needless to say, only a few vendors have the muscles and stamina that SAP has to tackle it in a do-it-yourself (DIY) manner.
I agree with Saugatuck’s assertion that cloud offerings are all about partnering across the ecosystem stack for most vendors. Thus, the more strategic thinkers will turn to established PaaS providers as the only sustainable cloud approach. While a number of small-to-midsize ISVs have begun to transition to the cloud using a variety of third-party tools and platforms, a few larger and well established ISVs have gone down the path of selecting a strategic PaaS partner such as Microsoft, Progress Software, or salesforce.com to help them accelerate their migration to the cloud.
A PaaS requires minimal procurement time and upfront cash outlay, given that it is priced by subscription rather than a large, upfront license fee. On the downside, because the standards are not yet established, all the aforementioned different cloud approaches are currently in use, posing some risk for a vendor who chooses a “wrong” (non future-proof) PaaS.
Infor and Microsoft in the Cloud, 24×7x365
Risky or not, cloud transition strategies for mainstream and established ISVs may increasingly be powered by leading and emerging PaaS platform players. During its annual Worldwide Partner Conference (WPC) 2010, Microsoft (as expected) continued to emphasize that it was embracing the Cloud as the core of its current and future strategy.
One key endorsement came from another large vendor that has been regarded as somewhat late to the cloud game. Namely, Infor made several key WPC 2010 announcements related to its evolving product development strategy. First and foremost, it announced the launch of Infor24, its blueprint for delivering cloud versions of its enterprise applications.
Key to this strategy is a strategic partnership with Microsoft (which was announced in late June 2010 and thoroughly analyzed in my recent blog series). Infor has decided to leverage a range of Microsoft solutions (i.e., Azure, SharePoint, Silverlight, SQL Server, etc.) as preferred technologies, as the vendor builds out the next-generation versions of its software portfolio.
The Microsoft alliance marks an important shift for Infor, as the vendor moves away from being technology-agnostic (which sounds good in theory, but has proven to be resource-intensive) to selecting a strategic technology partner to help move it into the cloud and hybrid environments. This shift is important for Infor so that it can both retain existing customers as well as attract new business with new solutions.
Thus, prior to Infor24, Infor announced Infor ION, a lightweight interoperability and collaboration layer previously referred to as Infor Open SOA. As pointed out in my recent blog series, the Infor ION framework is also the cornerstone of Infor’s hybrid cloud-to-on-premises software integration strategy.
With Infor ION, the vendor has essentially created a central data repository based on open data standards to help clients asynchronously exchange business documents. The collaboration business network part of ION will be enabling enterprises to conduct business transactions internally or externally, across both on-premises and cloud applications.
The idea is to provide out-of-the-box enablement for Infor cloud-based applications as well as a bridge to third-party cloud and on-premises applications. In addition, Infor ION adds business intelligence (BI) and business process management (BPM) layers on top of Infor’s core business application. For more information, see relevant blog posts from Infor’s Bruce Richardson and Ventana Research’s Mark Smith.
Part 2 of this blog series will provide a deeper analysis of Infor24, present some opposing views, and conclude with a question and answer (Q&A) session with Infor’s executives. In the meantime, what are your views, comments, opinions, etc. about Infor24 and the company’s recently revised product strategy? In general, do you think that Infor’s endorsement of the Windows Azure Platform indicates broader PaaS adoption by established ISVs, and the growing demand for cloud-based manufacturing ERP solutions?
Though I must admit that I have to read this article more critically once again before I totally subscribe to all the views, I felt that you are one interesting Analyst I always cherish reading and always look forward to working closely with you in the Cloud Computing Era!
Interesting article which touches not only on Software as a Service (SaaS) but also Infrastructure (IaaS) and Platform (PaaS) as a service.
Thanks for your compliments, Pradeep. Please feel free to provide some good natured critique as required :-)
Cloud Review Board - 5 governance responsibilities…
As enterprises adopt and adapt to cloud computing there is a need for a new governance role. This is on the basis that this shift is a business shift, with the potential to open up new business opportunities, and not just an IT infrastructure shift as …
quite useful info