Cloud BPM - All Thunder, No Rain ?
Cloud computing is definitely creating new opportunities for BPM product and service providers.
BPM vendors have already latched on to this opportunity. Late last year, Pegasystems announced its cloud enabled BPM PaaS with the promise of a platform that will not have the risks of security, performance etc that are usually associated with SaaS. Pega, being Pega, I am sure is coming up with something exciting and I am looking forward to the market’s reaction to it. Vitria too, had launched its VitriaCloud, which allows use of its M3O suite in the cloud model. Appian has its Appian Anywhere BPM PaaS offering and Fujitsu has introduced multi-tenanting capabilities into its Interstage BPM suite version 10 that allows its use as a SaaS. Other BPM vendors too have come up with some form of a BPM PaaS/SaaS offering.
Does it make sense for enterprises to go for these offerings? It definitely gives the smaller enterprises an opportunity to use the power of BPM, without having to make heavy initial investments. Even some of the larger enterprises that I have spoken to are approaching BPM very cautiously, and for them cloud BPM helps adopt BPM one step at a time, one process at a time. Compared to regular business applications, BPM applications have a greater need of participation of remote teams spread across geographies and cloud BPM enables greater collaboration than an on-premise implementation. However BPM, to go beyond just process modeling, needs to integrate with a multitude of applications within an enterprise and that may not be easy to do in the cloud model. Inter-enterprise processes that integrate with external systems are probably more suitable for running on an external cloud BPM platform. The concerns around data integration and data security in a cloud environment are also coming up in the BPM context. It is probably lesser of the issues though, considering that cloud based data integration solutions have been implemented successfully. (Infosys recently implemented a hub-in-the-cloud data integration solution for auto-dealers). May be the concept of an “on-premise cloud” will prove to be the best fit BPM for many large enterprises, but not for all. I think various flavors of Cloud BPM will be used and on many occasions Cloud BPM and traditional on-premise BPM may co-exist. The key to BPM success would be to identify the right model which would depend on answers to various questions e.g. which process, what is the scale, what are the integration touch points, what is the type of connectivity, what are the security requirements, what is the usage and most importantly what is the cost and pricing model!
Didn’t someone call cloud computing “complete gibberish”? I read that somewhere. Only time will tell if he was right. As far as Cloud BPM is concerned, I do find it promising. It definitely has some thunder in it. Will it deliver its rain or just disappear like a passing cloud? Whatever it is, I think BPM has some bright days ahead with some scattered clouds. Here is your local TV weatherman signing off!
Comments
Hi Arup. A nice writeup. And I totally agree with your views. In a nicely built SOA solution (bottom to top) within an enterprise, coarse granular services might be a result of orchestration/choreography of fine granular services (end systems exposed as services) and composite apps might be created using BPM (BPMN-BPEL) by orchestrating different coarse granular services. In such a layered solution, moving out BPM implementation to the cloud may not be attractive or of much benefit to the enterprise. But yeah for inter-enterprise process communication, across the globe, cloud BPM makes more sense. I personally view cloud as just a combination of Data/Computational Grid + Virtualization + Internet as platform. Hence I believe it can be positioned more at infrastructure/resource and deployment level... something like TIBCO ActiveMatrix Grid/container and then it's just a question of what piece of software (whether BPM, ESB, POJOs, Adapter, Rules Engine, portal, etc.) can be deployed and managed in cloud and can be used by clients inside and outside the cloud.
This thing may be more challenging to do for BPM software as there will be large number of communications happening to and from the cloud with external systems. Also BPM processes may be long running (sometimes for months)... in that case how will we effectively do the pricing model for resources. Some BPM software has 'dehydration' capabilities where long running processes are dehydrated to disk and activated sometime in the future. Can cloud BPM do that? Also how easy will it be to do monitoring, auditing, debugging, logging, reporting and management of business processes running in BPM cloud environment. I believe they may not be that straightforward and require more integration/collaboration with underlying cloud environment.
And as you mentioned the old age issues of security, performance, etc., will be always there. But again Governance and Risk has to play a key role in cloud implementations and something which enterprises will also focus on before taking any decisions. In a nutshell, I believe Cloud will inherently offer all the benefits of Web like distributed, on demand, scalability and also all disadvantages of web like security, reliability, etc. But obviously cloud is our "biggest bet" of bringing SOA, BPM and Web 2.0 technologies together and there will be more interesting usage of new business models like SaaS, etc as enterprise adopts cloud technologies with their existing SOA/BPM investments. But again, as you mentioned, different enterprises have to evaluate the right model for themselves based on different needs and evaluation parameters. I hope we will soon see some niche business models and success stories around the whole philosophy of cloud computing and cloud BPM, etc., in specific.
Posted by: Ankit Aggarwal | May 6, 2009 9:32 AM
Ankit,
Thanks for sharing your views in detail. Data Grid+Virtualization+Internet as a platform is a good way of looking at Cloud. Like most other big ideas in the world of integration, governance will probably play a greater role than the technology itself, in ensuring its success.
Posted by: Arup Raha | May 20, 2009 3:08 PM
Hi Arup, Ankit - One observation from my side is that you are probably from an EAI background. SOA is industrialized, nobody talks about SOA as a value add - SOA has become a commodity. And nobody talks about "external systems" anymore. We deal with business services, and they are standards based and secure. How is it different running within the enterprise or on a cloud? Some issues I see is real - Is the organization ready for cloud SOA? Is SOA industrialized in the organization? Is SOA used for business or integration? Are all services based on standards? Are services stand-alone? Above all - is the cloud SOA software ready for cloud?
The cloud SOA software should be a black box, developer friendly - and we never ever talk about performance, security and scalability. That's boring.
Posted by: Sreehari | June 12, 2009 2:31 PM