Adding Greater Spatial Capabilities to the Database

• 
• 

Details

By Matt Ball | 12 November 2010 | 1933

The spatial database performs a number of important functions for data integration and improved intelligence. These capabilities continue to expand current practice, and even push the possibilities. V1 Magazine editor Matt Ball spoke with Xavier Lopez, Director of Spatial and Semantic technologies at Oracle Corporation, about the company’s forays into 3D and real-time data, touching on the implications for cities as well as both CAD and GIS workflows.

The spatial database performs a number of important functions for data integration and improved intelligence. These capabilities continue to expand current practice, and even push the possibilities. V1 Magazine editor Matt Ball spoke with Xavier Lopez, Director of Spatial and Semantic technologies at Oracle Corporation, about the company’s forays into 3D and real-time data, touching on the implications for cities as well as both CAD and GIS workflows.

V1: Oracle has been busy on 3D data storage for some time. How long has it been, and what sparked it?

Lopez: Three years ago, Oracle Database 11g introduced a variety of 3D data types in the Spatial option. One of these is a native 3D (vector) type to manage city models. This type aligns with similar 3D vector formats like CityGML, Collada and even some CAD/BIM models. The business justification for providing support for 3D vector data in the database was the growing adoption and use of 3D city models across European and North American cities. A major driver is the decrease in development cost and increase in commercial availability of 3D models.

The second type of 3D support in Oracle Spatial is commonly referred to as point cloud data. Point cloud data are very large collections of three-dimensional points generated from laser scanners on airborne, mobile or stationary platforms. The points and associated metadata parameters record elements like: x, y, z, intensity, elevation, among others. Surveyors, planners and engineers now commonly use LiDAR to represent bare earth terrain, as-built structures, and to record the current 3D state of a development site

A common exchange format for LiDAR is the LAS file format which Oracle Spatial supports. By leveraging built-in partitioning, compression, and coordinate transformation support, the database is well suited to manage large point cloud workflows, in combination with other specialized pre and post processing software tools.

The third type of 3D data model available in Oracle Spatial were triangular irregular networks (TINs), also referred as terrain surfaces, or “two and a half D” data. As you know, representing 3D terrain surfaces is common use case for GIS and remote sensing applications.

Our goal for supporting these new 3D types is to ensure that our customers and partners can manage all their 3D spatial data in the same secure and scaleable way as they manage their 2D and business data.

V1: Primarily it sounds like advances in data capture have spurred the 3D development. Is there also a look at fusing the data types?

Lopez: Spatial data types are becoming richer as the sensors gather new inputs, and as the underlying 3D models refine and standardize the underlying metadata and semantics. Some of this evolution will include a new generation of hybrid spatial types that incorporate 2D, 3D and temporal elements. Today, however, fusion of information is undertaken by the applications themselves. Most of the leading GIS and web mapping tools can fuse discrete 2D vectors, rasters, and 3D data into a spatially consistent and coherent visual representation. This type of fusion via map rendering is commonly done in most GIS and 3D globe viewing tools.

From a database perspective, we want to ensure that customers can securely manage these data types, while ensuring the scalability and performance for environments ranging from the desktop to the cloud.

From the database perspective, it is important for us to keep up with this rapid evolution taking place in the output of 3D laser scanners and photgrammetric tools. We are also committed to ensuring that 3D data stored in an Oracle Database can be accessed, analyzed and viewed from a variety of commercial and open source tools. As a result, our customers benefit from the same consistent data management environment for all their spatial and business data. This common data management environment helps organizations consolidate the management of costly stove-piped systems.

V1: I imagine the data storage growth chart is amazing. Are there any figures that you chart in terms of the demand for database storage?

Lopez: For the last decade, data growth, along with storage and management, has been a top 10 technology trend in IT. Simply look at recent 2009 and 2010 industry surveys by ComputerWorld, Gartner, and IDC for confirmation. With geospatial data, the growth trends are compounded by the complexity of new spatial types, and the continuous flow of various sensor-based network data used in mobile computing.

LiDAR, in particular, is making significant ripples in the geospatial market segment. Observers in the surveying, engineering, architecture, forensic environment will notice that 3D laser scanning has reached a price-point where it’s faster and more effective then traditional surveying alone. Moreover, the derivative products from the surveying project not only improve the deliverable, but result in additional high-value data products. There is a classic spill-over effect from LiDAR transforming not only the surveying community, also the various disciplines that use GIS and remote sensing data.

V1: The 3D capability really starts to merge the worlds of GIS and CAD, and there are certainly some software visualization efforts to bridge that gap. Do you see that challenge as more of a user-side issue or a vendor-side issue?

Lopez: We need to recognize there are both user and vendor challenges. On the user side, engineers and architects, for example, have traditionally been trained in the use of CAD tools for their design work. However, new requirements on their tasks require them to take into account new data inputs and new tools into the professional work practices.

In the software tools arena, some vendors are trying to create these bridges: CAD to GIS; GIS to CAD, CAD to simulation, GIS to simulation, and so on. Others are trying to incorporate these alternative features natively. Standard data specifications can help ensure that these CAD and GIS data types are interoperable and more broadly used. Undoubtedly, there is plenty of room for improvement in all of the software tools currently on the market.

V1: It seems like a 3D model that can incorporate a lot of different inputs specific to a lot of different domains is the ultimate goal here. Is there going to be a convergence across a lot of different software vendors and tools? Do you think we’ll get to the point where the model is the interface?

Lopez: The 2D map paradigm is an effective metaphor for a variety of engineering, analysis and reporting processes. However, adding a third (or even a fourth dimension) can vastly improve data quality and usability of many applications.

The 3D model is powerful in a number of ways – as a rich visual representation, as interface for interacting with complex spatial features, and logical model for representing feature relations to support non-visual analysis and reporting

V1: It gets really interesting on the city scale, with tools to manage all the inputs. The city itself is a very complex model where system science could be employed to have a better understanding about what’s going on, beyond the simple inputs we have now. Are you working at the city scale as well?

Lopez: Supporting enterprise spatial systems is Oracle’s forte. We’ve developed our spatial database to provide urban planners, transportation engineers, utility operators, first responders, and others with a solid software platform for managing a real-time, digital city infrastructure. The complexity of data types, variety of data feeds and volume of structured and unstructured data in this scenario is a challenge. For example, applications need support for 3D buildings with textures, 2D building footprints, road network topology, land records, and on and on. Oracle continues to innovate, by offering new platforms such as the Exadata and Exalogic database and middleware products to support the ultra-high performance, security, and scalability requirements of public sector cloud infrastructures.

V1: It’s interesting to also think about the incorporation of sensors as well into these networks, to create intelligent infrastructure. IBM is working strong on their Smarter Cities initiative to brand this idea of a smarter planet. The focus is on the integration of these different pieces for that holistic view. Are there initiatives at Oracle that bridge across different domains to create this sort of environment?

Lopez: From an industry perspective, Oracle has been active in various Smart Grid activities in the utilities sector. Other industries that have warranted our involvement include intelligent transportation systems (ITS) and a range of defense and intelligence initiatives. The idea of Smart Cities is exciting, but as you can image there are huge challenges in gaining acceptance from the various data stewards to share and integrate their information in support of public services. While there is an important technology aspects to this problem, I tend to see this more of an organizational challenge for now.

V1: There are various cities from scratch initiatives in Asia, such as South Korea’s U-City initiative where the cities systems are being design for transparency and centralized control. There’s also the look at the city as an urban metabolism, where the systems feed each other for a better idea of the city as a whole. I don’t know how much of that is 3D, or the degree and level that sensors are incorporated into these models, but it does marka new degree of system integration.

Lopez: Modeling the “city as organism” an interesting concept, but not necessarily new. For urban planners there is a tradition of modeling the city is as an interactive web of networks. This is similar to the perspective of earth systems scientists that attempt to model the Earth as a large complex ecosystem. In each case, high quality reference data and real-time sensor feeds are key to the monitoring, reporting and simulation activities. Clearly, a powerful and flexible IT platforms that can handle the large volume of rich spatial data types and real time feeds becomes critical.

V1: It seems that there’s a bit more urgency now, given different impacts of global change. It’s interesting to think of such things as ecosystem services for a valuation of what the natural environment provides.

 

Lopez: That’s an interesting observation. We’ve also noticed the adoption of geospatial processing and visualization in applications ranging from monitoring CO2 sequestration to assessing the optimal location of wind and solar farms. And of course, users in the area of natural resource planning continue to innovate with the adoption of spatial tools, sensor devices and a variety of temporal scales. It’s an interesting time for innovators and entrepreneurs to start combining these technologies and content in new and meaningful ways.

V1: All of the web services applications that are cropping up are mainly serving data rather than becoming the sort of platform for interaction and collaboration. Piecing it all together and getting a holistic understanding of what’s going on, requires some serious data management and integration.

Lopez: Data management is important because you have a lot of this data that isn’t necessarily transient or streaming, and needs to be stored. A lot of the streaming data needs to be stored as well, but not all of it, because it may not be practical. That said, the challenge is much greater than data management. The other challenge is data integration. Hence, middleware becomes a critical technology layer to enable data and web service integration to take place.

Middleware, such as the Oracle WebLogic Server, provides a web services infrastructure that enables the filtering, transformation and integration of different web services feeds from different sources. In particular, Oracle’s complex event processing (CEP) engine is designed to handle such complex sensor workflows in combination with geoprocessing.

CEP is commonly used in air traffic control and mobile tracking applications, where location tracks are streaming in with position and state of vehicles and planes. The Oracle CEP engine has been uniquely enhanced with built-in spatial business logic, such as geo-fencing, within distance operations, and web map visualization to support a wide range of mobile tracking applications.

V1: We used to talk about the stovepipes and silos of different systems within an enterprise, and now we’re talking about domain barriers that stand in the way of a more holistic view.

Lopez: It’s been my view that the diversity and heterogeneity of IT systems reflect the legacy, objectives, and constraints inherent in organizations. These organizations can range from a small work group, to a department, or an enterprise. However, today’s realities are forcing organizations -- both in the public and private sector -- to respond to change, to new mandates, new opportunities, and to shifting public expectations. Given the difficulty of organizations to effectively respond to this dynamic business environment, many are relying on technology to overcome the brittleness inherent in organizations, hence, the increased interest in web service specifications and standards.

V1: Are there any developments in the visualization space that are interesting to you?

Lopez: I’m delighted to see the GIS community taking 3D visualization seriously; and the CAD industry taking simulation and analysis more seriously. New standards in the visualization and modeling space – CityGML for city models, IFC for building information models (BIM), and X3D for 3D objects are important foundations for interoperable visual tools. The work going on with CityGML and BIM is very important since it leads to smarter 3d models that can integrate with enterprise business systems. The adoption of this model based approach also helps move the construction and engineering fields to a higher level of data efficiency, re-use, and productivity.

A model-based approach to design and operations also enables design data to readily integrate into the of asset and property management workflows. This example, illustrates the evolution of 3D models from “design projects” output to “operational workflows” input that flows with operational lifespan of a particular building or asset. In the virtual city use case, these workflows are ongoing and manage the large volume of 3d vector solid models and associated textures, 2D vectors, imagery, point clouds and video feeds. Moreover, these spatial features are keyed to additional attributes and business systems.

V1: It’s great to see on the CAD practitioners move toward larger models. On the GIS side, Esri is talking about GeoDesign, and they’ve incorporated some 3D capabilities in their latest release. It looks like there’s a coming together on both sides.

Lopez: Yes, there is a convergence in the feature requirements of both CAD and GIS tools. We also observe new requirements for simulation, animation, collaboration, and complex event processing. These requirements create new opportunities for the software community. It will be interesting to see what new tools emerge over the next five years to address these exciting requirements.