Without solutions to the problems of data sharing, Grids and other forms of agile IT will not be able to support key enterprise applications. This will prevent enterprises from fully exploiting the benefits of consolidating IT resources. Instead, new silos will be created, with key enterprise applications separated from those that are able to run safely on Grids. The result will be continued IT over-provisioning and an inability to react in an agile but robust manner to changes in demand for enterprise applications. It will also prevent the enterprise from responding rapidly to new business demands that require data to be shared across those applications.
A key question for Grid providers is how to evolve their infrastructures to embrace transactional enterprise applications. New solutions are required to achieve this. Based on our experiences in managing more static forms of data sharing in enterprise applications over the past two decades, we believe that these new infrastructures must be able to:
We have not yet seen these features in commercial Grid infrastructures. Many 'compute grid' infrastructures do offer fault-recovery through restarting the applications - not sufficient for enterprise applications that share data. Many 'data grid' solutions focus on data coherency without providing scoping mechanisms for data consistency - not sufficient for enterprise applications. Such solutions could result in inconsistent enterprise data being exposed to other applications and so propagated through the organization, with potentially disastrous consequences.
It is, therefore, essential that CIOs, IT managers and buyers are aware of the potential problems that may affect their enterprise applications as they move towards a more agile IT infrastructure. Enterprises should consider these issues when evaluating the proposed solutions offered by Grid technology vendors.
Arjuna is currently prototyping a solution that embodies our ideas on how to support reliable data sharing in an agile infrastructure.