The small relationship between PAAs and large data

As the Internet of things and the deepening of large data applications, the value of perceived data has become more and more well known through various devices. The integrated use of perceptual data embodying the physical world and real-time operation can fully excavate the value of data and play a key role in solving hot spots such as traffic jams, environmental pollution and road network layout.

A team of Li, Dinwilong and Zhaozhou, from the Cloud Computing Research Center at the Northern Polytechnical University, "Stroll the Cloud", leveraging Microsoft's Windows Server 2012 Technology Advantage and sophisticated package solution for infrastructure virtualization, to build a platform for perceived data hosting and application services, To enable tenants to quickly and easily develop and deploy applications, the realization of traffic-aware data integration, calculation, storage class purposes, reduce the cost of data-related applications, and meet the needs of data/application hosting and operational dimensions.

Breaking the information island of intelligent traffic

Zhaozhou, vice director of cloud Computing Research Center, North China University of Technology the original intention of building a perceptual data hosting and application service platform is to solve the contradiction between the need for the application of Internet applications and the relative lack of common infrastructure, and to realize the deep integration of data.

Such contradictions are particularly evident in the field of transportation. Traffic perception data has the characteristics of mass, high concurrency, continuous and high speed. According to the previous practice, if the data volume increased or application increased, the pipeline department will be the infrastructure of vertical expansion, the purchase of equipment such as servers. This is not only expensive, but the increase in the size of the server after a certain number of vertical expansion is becoming more difficult. From the actual point of view, the amount of data increased very quickly.

For example, the construction of the control system of city trip vehicle based on license plate recognition The initial verification phase of the project may only be connected to dozens of cameras with license plate recognition, the number of cameras accessed after the feasibility verification is likely to expand to about 1000, eventually in a large city may be connected to nearly million cameras. These camera data and accompanying vehicle photos will be assembled into the system for analysis, according to the thousands of cameras, each day may accumulate more than 30 million, nearly 10TB of vehicle identification data. As the volume of data increases dramatically, the ability to store and compute infrastructure becomes a bottleneck in the construction of the application system. According to the traditional system architecture, some infrastructures may be less efficient in some cases. This is because the current camera distributed on the road is usually used to serve different systems, different applications, the peak of the application load is not the same. Moreover, these different applications may form more islands of information, and can not realize the cross application and cross departments sharing of basic traffic data such as license plate recognition.

If the infrastructure currently serving different applications can be integrated and shared through virtualization, while in a multi-tenant manner, supporting intensive storage, shared management and related applications of perceived data across different business areas, it will not only be a good solution to infrastructure expansion, It resources optimization configuration and IT system to reduce operational costs of the problem, but also can achieve the basic sense of the sharing of data, reduce the depth of the development and utilization of perceived data, promote business innovation based on perceptual data, truly realize the transformation of data value to business value and social benefit in large data age.

It is in this context that the roaming cloud builds a platform for perceived data hosting and application services based on Microsoft Windows Server 2012.

Virtualization is just one aspect

In an interview, the reporter learned that the choice of Microsoft Windows Server 2012 in the cloud was due to Microsoft's ability to provide full virtualization capabilities including IaaS and PAAs layers, on the other hand, because Microsoft Windows 2012 has a very good user experience in the configuration interface, Especially makes installation, deployment very convenient and easy.

In the IaaS layer, walk the cloud select the hypervisor based operating system Windows Server 2012, Management components system Center. This is primarily due to the three considerations of extensibility, compatibility, and usability.

As the perceived data hosting and application services platform is concerned with resource consolidation and optimal configuration, roaming the cloud seeks to pool the capabilities of the processor, network connectivity, and storage resources on the product selection, providing the tenant with resources on demand. System Center 2012 provides end-to-end management of the infrastructure across physical, virtual, and cloud environments, with unified management in all locations. Hyper-V ensures uninterrupted migration of virtual machines between different host hosts, and a high availability solution that can be fault-tolerant to ensure that cloud-based services are uninterrupted. In addition, Windows Server 2012 delivers more scalable, resilient cloud solutions that are truly optimized for the cloud. Windows Server 2012 and Hyper-V support a cluster with 64 nodes, up to 4,000 virtual machines, and a relatively small amount of physical hardware to support and deliver greater capacity for IT services. In addition, the cloud walks through the public API of power Shell 3.0 to facilitate application developers to invoke interfaces on the PAAs platform as needed.

In the PAAs layer, stroll cloud Select Application Server IIS, directory service Active Directory, cache service BranchCache, and development environment visual Studio. These can provide efficient Web application hosting environment, layered distributed tenant authorization management, and realize the management of virtual machines and scheduling of computing tasks.

Walking Cloud team member Lee told reporters that the template (including the virtual machine template and the application template) is the most critical and complex part of the entire solution build. It is these templates that enable users to quickly develop and deploy new applications.

A virtual machine template is a combination of "mirror-configuration-description information" created by an administrator on a platform to mirror-bound CPUs, memory, and so on. The application template is a combination of virtual machine templates, created by the administrator, for the tenant to instantiate the appropriate virtual machines for deployment applications. In addition, when application templates do not meet the requirements, tenants can also append virtual machines to existing applications by instantiating or manually creating virtual machines.

For example, in a traffic project that walks the cloud, a typical traffic flow data business computing application, deployment and operation requires three types of virtual machines: Communication Server, compute server and database server. Through the platform, the administrator can implement the design to provide an application template containing the three types of virtual machine templates, for the relevant transport Department of business tenants to instantiate the relevant virtual machines, to achieve the deployment of the corresponding application purposes.

Walk the Cloud team member Dinwilong, demonstrating the platform's four key competencies. Firstly, it supports the fast access of different kinds of perceptual data streams, and provides real-time receiving and routing distribution of data with high concurrency and high swallowing capacity. Secondly, based on virtualization technology, the management and dispatch of basic resources are realized, and the problem of operation and maintenance of complex systems is solved. Third, A transparent parallel computing interface and cluster computing environment based on large-scale perceptual data are provided. The Forth is to realize the fusion management of multi-source data and related business data, and provide the service interface of data sharing and integration.

Laying the groundwork for industry informatization integration based on perceptual data

The sensing data hosting and application service platform realizes the application-driven virtualization resource management and scheduling of IaaS layer. On the one hand, the application-centric resource Virtualization management tool is constructed, which realizes the organization and management of the IAAS layer virtualization resources from the perspective of perceptual data and application hosting and transparent deployment. On the other hand, it realizes the resource virtualization Scheduler using state-aware, and monitors the data storage and application running state, Automate the adjustment of the IaaS layer virtualization resources.

Zhaozhou points out that the value of the perceived data hosting and application service platform is mainly embodied in the following three aspects:

First of all, reduce the cost and cost of application development under the background of IoT and large data. Even small and medium sized software companies can develop intelligent/intelligent applications based on mass-aware data through a PAAs platform such as the perceptual data hosting and application service platform.

Second, significantly reduce the deployment costs and operational costs of the intelligent/Intelligent Application system described above. The perceived data hosting and application service platform, through transparent management and optimal configuration of the infrastructure, reduces the deployment cost of related applications and reduces the complexity of system deployments through virtualization. In addition, the perceived data hosting and application service platform greatly reduces the complexity of operation and data expansion and migration, and reduces the operational cost by means of automation such as system monitoring and load sensing.

Finally, the technology provides potential support for the centralized management and integrated application of perceived data (such as traffic perception data) in future industries. Zhaozhou that the perceptual data hosting and application service platform provides the possibility for the future industry Informatization integration based on perceptual data. In the traffic field, for example, in all kinds of traffic-aware data (logical or physical), the traffic-aware data which originally served for a single application can be shared by different applications. For example, the data generated by the camera used to limit the traffic can also be used to deal with speeding vehicles, real-time road traffic calculation and other services, and apply to related systems. Moreover, by further integration of the perceptual data of different industries at a higher level, it can also be used in the construction of intelligent cities, such as urban planning, urban management, emergency monitoring and early warning, public information service, etc.

翟爽, senior architect of the Windows Server 2012 Cloud Creation Competition judges, Microsoft development Tools and Platform division, is deeply impressed with the idea of walking the cloud team. "The construction of intelligent city needs not only the construction of infrastructure, but also the innovation of application level." Based on products such as Windows Server 2012, the roaming cloud team enables tenants to quickly and transparently deploy access, computing, storage-class applications such as a traffic-aware data, reduce the cost of sensing data and related applications, and meet application/data hosting and operational requirements. "Zhang Lijun said.