Issue link: https://read.dmtmag.com/i/50311
to generate easy-to-understand reports, maps and graphics for decision-making processes. The latter group is the Viewers, who represent most of the technology users at large. The Viewers require high-end capabilities that allow them to quickly access detailed data and analyses that are easy to under- stand for their business-intelligence purposes. These easy, yet comprehensive, reports and analyses are, in fact, created by Doers behind the cloud. Ron Lake: Cloud-based computing clearly is here to stay. However, I doubt that all applications will go that route, at least not in the foreseeable future. Where I can see clouds dominating is those applications that need to be shared across a broad spectrum of users, including users with limited IT resources. This will include a lot of government information sys- tems such as systems for the issuance, monitoring and regulation of permits, licenses, leases, etc. The economics of cloud-based infrastructure, and the ability to reach all users (e.g., a small municipal- ity), will simply be too significant to ignore. Concerns about data security and privacy for such environ- ments will have to be dealt with, however, that can be readily achieved by government-owned clouds or simply stricter regulation of private-sector clouds. Very-high-end applications (e.g., Google Earth) already are in the cloud, but I anticipate that many large enterprises will keep some of their key applications inhouse, if only in their own cloud. David S. Linden: Many geotech- nology applications fit very well into this model, particularly those with small- to medium-size datasets run on networks provid- ing Internet or intranet access. However, many other geotechnol- ogy applications don't fit this model, because their datasets are too large or they run on small, isolated networks. The latter case describes the environment my applications typically run under. I work with large raster-based datasets that include high-resolution imagery as well as high-density digital elevation models. There's really no efficient way to distribute these data in a cloud-computing model. Furthermore, I often run applications on isolated networks that have only a handful of computers on a local-area network. There's simply no access to a cloud. So although cloud-computing does stand to benefit many geotechnology applications, it's not a panacea of computing power for all. Dale Lutz: There's no doubt that cloud-computing architectures are revolutionizing the IT land- scape, and geospatial certainly is no exception. Everywhere I look, I see deployment of geospatial services into the cloud. The "fly in the ointment," how- ever, is the always increasing data volumes that geospatial requires. Where those huge datasets can naturally be made to reside in the cloud to begin with, then there's no problem moving all of the processing work into the cloud and enjoying its scalability. However, where those datasets are large, change frequently and need to be moved up to the cloud, then the fit isn't so good. Sadly, there are many cases of this nature. As a result, and with most things, there's no "silver bullet." The future will consist of a mix of cloud-based services, intermingled with locally hosted solutions, with what's selected depending on the nature of the problems as well as the volatility and locality of the data. In the end, the applications need to reside with the data, and that will prevent a wholesale migration to the cloud. Carey Mann: A world where we completely abandon our applica- tions and storage to remote servers owned by a handful of companies seems unlikely. What will emerge is a hybrid environment. Geospatial applications are characterized by many of the stress points normally thought to benefit from cloud architectures. Collaborative workflows between orga- nizations may require a demilitarized zone suited to the cloud. So will processes that are computation- ally intensive or subject to load imbalance, where the memory and processing of a cloud service could be leveraged. Storage of large, static and non-strategic data- sets as well as format and data transformation services could be attractive. Although some appli- cations might be completely cloud-centric, it's likely most will combine on-premise implementations incorporating cloud technologies, embracing the hybrid model. DECEMBER 2O11 / WWW . GEOPLA CE . C O M 23