This two-part article about "the cloud" is from Joel Campbell and Amy Zeller of ERDAS. Part one described how technology maturation - from the early days of "time-sharing" on big computers to the current "pay-as-you-go" model - has brought us logically to the era of cloud computing. In part two, they discuss the benefits of cloud computing and various implementation models.
Benefits of the Cloud to Geospatial Organizations
Geospatial organizations need to securely make volumes of geospatial data and services available to their customers. They require solutions to easily store and manage data, and also supply rich sets of Web services that enable not only data delivery but also a variety of unique functions such as catalog queries, data and coordinate transformations, vector editing, mapping capabilities, geoprocesses and much more. For isolated events, infrastructure must be elastic to accommodate sudden increases in demand so that strain does not become overwhelming, and to avoid unnecessary capital expenditures on hardware, software and services.
Most organizations, intent on accomplishing all of the above but still operating on limited resources, are more interested in growing their geospatial human resources than adding to IT head counts, IT expertise or expanding IT infrastructure in-house. Therefore, geospatial organizations have developed a keen interest in powerful, secure and reliable off-site infrastructure solutions that are cost-effective and rapidly scalable. Hence their interest in cloud computing.
Benefits of cloud computing to the geospatial industry include:
- Secure, turnkey solutions can be running in a matter of hours
- Solution delivery is simplified: installation and configuration becomes a known, repeatable process
- Reduces support calls
- Allows more time to focus on R&D
- Improved security due to focused resources and protection systems
- Low barriers to entry: quick and easy to get up and running on the cloud
- Eliminates the need for in-house, physical resources, new IT head count and expertise
- Scales instantly to meet spikes in demand for both services and storage
- Organization pays only for actual resource consumption and avoids unnecessary capital expenses on hardware, software and other services
- Pay as you go, terminate when you want
- Shared infrastructure costs and low management overhead = savings in your pocket
Regardless of these benefits, however, significant fears still exist regarding the use of the cloud. Some of the biggest challenges to adoption revolve around fears of data security, availability and fluctuating costs. None of these fears are new, though; they've actually been around as long as off-site data hosting providers have been an infrastructure option.
Layers of the Cloud
Because "the cloud" is a term that is broadly used throughout the IT industry, it is important to understand its manifestation within the geospatial space. Breaking it down, we see four main cloud models:
- Big Cloud Infrastructure: These are the rapidly elastic, pay-as-you-go cloud services offered by companies like Amazon EC2 and Microsoft Azure. These companies focus on providing you the shortest route to accessing and scaling infrastructure resources. These companies offer a more generic cloud offering with limited support for the implementation.
- Add-ons to the Big Cloud Infrastructure: Companies are now emerging that implement industry-specific offerings on top of a big cloud provider. Some examples in the geospatial industry include WeoGeo and SpatialCloud, which both provide specialized solutions on top of Amazon EC2. WeoGeo offers geospatial data storage and transformation capabilities, while SpatialCloud enables application developers and content providers to store and distribute geodata. Delivering these services on top of Amazon EC2 allows these companies to offer flexible and scalable infrastructure and storage capabilities as the foundation for their specialized solutions.
- Niche Cloud Providers: These are smaller cloud providers who build their cloud from the ground up to satisfy particular industry requirements. A great example of this is our partner, Skygone Cloud. Skygone operates its own cloud powered by state-of-the-art data centers. Skygone's unique offering is that it employs highly skilled GIS experts who are strictly focused on a solution-driven approach to cloud-based GIS implementations. Skygone Cloud staff is made up of many technical and business leaders hailing from some of the leading GIS software and service companies in the world. Via the cloud, they seek to transform the way the GIS industry builds infrastructure solutions to support its mapping operations.
- On-premise/Private Cloud: This is basically building your own internal data center to run in an efficient, cloud-like manner. Many organizations want to harness the flexibility of the cloud, yet their fears of public clouds include loss of security, privacy and reliability. The solution is to create a cloud infrastructure within the IT framework of their own organization. Some companies that offer on-site implementation of an on-premise cloud include Skygone and Appistry.
Transitioning Existing Offerings to the Cloud
Many geospatial software providers are equipped to easily transition their portfolio of products into cloud based solutions, including systems for data management and delivery, geoprocessing, vector management and editing, location-based services, photogrammetry and general mapping capabilities. For the broader geospatial industry, we see a need for the following cloud-based offerings:
- Hosted Cloud-based Solutions: Customers may want to purchase and use geospatial solutions, without investing in infrastructure or running the software in-house. Instead, they choose to kick-off and run applications with no long-term contracts in an off-site, scalable environment. Customers can either bring an existing software license or rent the software license along with the infrastructure.
- Infrastructure On-demand: In this scenario, customers already have the geospatial software they need, but not the computing power. For example, a county may be using ERDAS IMAGINE to generate an orthomosaic. This operation takes a lot of processing power, but it is only run once a year. Consequently, the county may not want to invest in major systems for this one project, nor does it want to wait a week while the processes run over light-powered, in-house systems. Instead of sending the job to internal systems, an option would be available within the IMAGINE user interface to send the process to the cloud. Via this option, the entire process would be forwarded to powerful, off-site cloud CPUs. This is truly scalable infrastructure on demand, easily supported by big cloud infrastructure where everything is pay-as-you-go and elastic. Development of this capability, however, will require significant R&D to understand such concepts as how to potentially connect internal network data to CPUs on the cloud.
- Software as a Service (SaaS): Some customers do not want to simply buy geospatial software or computer power, but want to buy a full-service, customized, managed geospatial solution. In this scenario, the customer is not only relieved of infrastructure ownership, but also relieved of solution ownership and management. Similar to a Salesforce.com approach, in this case the geospatial vendor would completely own and manage the solution for the customer.
First Offerings to Early Adopters of the Cloud
The cloud computing game has just begun. Now is the prime time for early adopters who can see beyond the hype and fear and recognize the value of cloud-based geospatial solutions.
We see that utilization of cloud-based solutions is where the industry is moving, and we've adopted this first tier as a starting point. The reality is that the cloud offers geospatial providers a "silver lining" amidst a perfect storm of requirements for authoring, managing and delivering geospatial data. This is truly an exciting time for our industry and we look forward to offering a variety of solutions in this space.