What is a photonic grid? Come up with a more accurate answer

PopeЪ puff (E-science) [1]. It combines scientific instruments, high-performance computers, distributed databases, sensors, and remote devices located in different geographic locations to solve complex scientific problems, such as global climate simulation, high-energy physics, mapping of genetic profiles, nuclear test simulation, development of new drugs, virtual expert consultation, large-scale information and decision support systems. Grid technology enables people to *** enjoy computing resources, storage resources and related services, so it plays a vital role in scientific research programs and industrial development in astronomy, aerospace, transportation, automobile manufacturing, meteorology, steel generation, nuclear reactors and many other fields. In grid applications, sensors, remote devices, high-performance computers and visualization devices need to transmit massive data of terabyte or even petabyte magnitude in real time between them, and the network occupies a very important position in grid applications. The traditional Internet cannot provide high-speed transmission of massive data with low latency, and its best-effort service mode cannot meet the QoS requirements of users. Therefore, the grid applications built on the traditional Internet have many limitations, such as slow data transmission, poor reliability, poor user interactivity, the use of interface is not friendly enough, etc., greatly affecting the efficiency of the application system. Photonic grid (opticalgrid) is an emerging technology developed under the above background in recent years [2~3]. Its basic idea is to interconnect high-performance parallel computers, computer clusters, large-scale storage devices, high-definition display devices, large-scale scientific instruments, and various types of personal computers, servers, etc., which are distributed in different geographic locations through optical networks. Since optical networks have large bandwidth, high transparency, low latency, low cost, high reliability and dynamic bandwidth adjustment capabilities, the photonic grid can also provide rapid transmission of massive data, high reliability management and flexible scheduling and control of resources for grid applications while meeting the needs of users*** enjoying grid information resources. The implementation of Photonic Grid network can break through the network bottleneck problem in Grid applications, enabling Grid users to keep in sync with high-end computing resources and maintain satisfactory interactive functions, thus accelerating the scientific research process in application areas and promoting the development of related industries. On the other hand, the implementation of Photonic Grid enables the owners of high-performance computing resources, storage resources and scientific instruments to expand the application market more effectively and improve the utilization rate of resources. It can be seen that the photonic grid is a feasible technology to make grid applications truly practical.2 Background of Photonic Grid In recent years, with the continuous development of large-scale scientific computing applications, the requirements for computer processing power, storage capacity and high-performance visualization are increasing. Computer processing or storage capacity is subject to technical and cost constraints, and it is neither economical nor practical to equip each user with high-performance computing, storage and visualization equipment. A feasible solution is to assign computation and storage tasks to different computers, and to realize large-scale scientific computation and visualization applications by *** enjoying the computation, storage and visualization resources of different research institutions. This approach can effectively save costs and improve the utilization of resources. At the same time, the complexity of today's scientific computing problems is increasing, which requires scientists from different fields and countries to *** collaborate in order to achieve breakthrough results. Therefore, it is necessary to build a high-speed network to link these scientific researchers, high-performance computing and storage devices, high-precision instruments and visualization equipment, and to achieve efficient transmission of massive amounts of data between different geographic locations. The above applications have led to an increasing demand for network connectivity and bandwidth. The extensive laying and wide application of optical fiber and optical network transmission equipment provides the possibility of interconnecting high-performance computers, large-scale storage devices, high-definition display devices and large-scale scientific instruments. At present, in 10Gbit / s and higher rates, compared with IP switches, optical switches have lower power consumption and cost. Optical networks can provide low-cost, high-bandwidth, high-reliability optical connections, which have been accepted by the vast majority of research organizations and even some individual users. Photonic Grid, which is created in the above context, associates end-users, computation, storage, and other resources through optical networks, thus realizing high-speed transmission of large amounts of data remotely.3 Key Issues in Photonic Grid ResearchPhotonic Grid is not the same as simply using an optical network to provide big data transmission. To effectively support grid applications, traditional optical communication networks and grid technologies face a series of challenges. First, to support grid applications, transmission bandwidth on the order of Mbit/s to Tbit/s needs to be provided for a large number of users and end devices. Users' requests for bandwidth are characterized by burstiness, parallelism, large scale, and coexistence of multiple granularities, while the bandwidth resources of optical networks and the computational and storage resources of grids are limited. Obviously, it is neither economical nor practical to provide a dedicated optical path for each user task. Therefore, the optical communication system needs to support different types, multi-granularity, sudden bandwidth demand, with the ability to distribute bandwidth on demand; provide multicast and broadcast capabilities; at the same time, the system in order to meet the application requirements, but also need to provide users or applications with self-organization, self-management and self-control of the distributed network resources, to support the establishment of a flexible, fast channel. Secondly, grid applications are different from point-to-point communication services on communication networks, which are characterized by distributed, multi-task flow work, where multiple tasks can be assigned to different computing resources to run in parallel, and different task allocation methods will lead to different optical network resource allocation methods. Even if the computational resource allocation scheme is determined, the optical network resources will have different scheduling schemes because there can be different routing choices between the optical channel source and host node pairs. In turn, different task allocation methods will lead to different task completion times. Therefore, to efficiently complete a given service under the given constraints, the system must support large-scale distributed parallel network services, must reasonably describe the interrelationships among business processes, and must collaboratively schedule computational resources and optical network resources in a completely new way, or else it will directly lead to the reduction of system operation efficiency and resource utilization. Moreover, in the process of resource discovery and task scheduling, Grid computing usually does not consider the limitation and availability of network resources, and lacks a discovery mechanism to obtain available network resource information from the network. In practical applications, network resources are an important factor that affects system efficiency and application efficacy. Therefore, it is necessary to find a new resource description, resource discovery and resource update mechanism to realize the unified management and reasonable utilization of computing and network resources. Finally, the multi-service flow and large data volume characteristics of grid applications require communication networks to have higher security and data correctness assurance. Although the grid has a certain fault tolerance mechanism and the network has a certain protection/recovery capability, how to realize a higher level of system fault tolerance through the cooperative operation of the optical network and the grid according to the user's QoS demand, in order to ensure the security of the network and the security of the interface between the grid users and the communication network, is also a problem that needs to be solved. Aiming at the above key issues, domestic and foreign research organizations and related scholars have carried out research on photonic grids and their applications focusing on the following aspects. -Photonic Grid architecture and realization technology: focusing on the technology of constructing photonic grids, the basic composition and functions of photonic grids, the interrelationships of the components of photonic grids, the ways or methods of integrating the components, and the interrelationships between them and grid applications. -Control and management protocols: focusing on the control and management mechanisms of photonic grids, including user network interfaces, computational resource invocation and control mechanisms, dynamic invocation and adjustment of optical network burst bandwidth, signaling and routing protocols, inter-domain and inter-layer control protocols, and interface technologies and implementation methods of photonic grid middleware. -Photonic Grid resource discovery and scheduling mechanism: Focus on the description, registration, release, update, service deployment, resource discovery and resource scheduling mechanism of Grid information resources and optical network resources in Photonic Grid environment, and on this basis, study the cooperative and optimized scheduling mechanism of Grid information resources and optical network resources under different working modes, implementation algorithm, and performance index analysis. Photonic Grid Fault Tolerance and Secure Access Mechanism: Focusing on the research of photonic grid authority management mechanism, user authentication technology, and security and authority management technology for cross-domain scheduling, research on how to set up different levels of fault tolerance strategies in the event of fiber link interruption, equipment node failure, server downtime, or interruption of service programs in the photonic grid, so as to safeguard the accuracy and timeliness of data transmission, and at the same time, make the user The system failure can not be detected to meet the QoS requirements of different users. Business Model and Application Experiment: Focus on the classification and organization of business types under various grid application models, formulate different business level mechanisms for different types of business according to the QoS requirements of users, give the description methods of business workflow under different types and levels, and provide a visualization tool to assist users in defining the process and generating the description files, and on the basis of which, we aim at high-performance computing and visualization, large-scale collaborative computing, and large-scale collaborative work, and provide a visualization tool for users to define the process and generate the description files. On this basis, we discuss the realization technology, application process and development prospect of photonic grid under multi-service application mode for typical applications such as high-performance computing and visualization, large-scale collaborative design and real-time data transmission. It can be seen that building a new type of network architecture, integrating network, grid information resources and services, and realizing the cooperative management of end users, network resources and grid information resources, there are many problems to be further explored in both theoretical research and practical application.4 Photonic Grid Research ProgressAt present, relevant organizations at home and abroad have carried out a series of research work in the field of photonic grid, and the representative research programs or projects include: - OptiPuter project [4] in the United States, which interconnects computer cluster systems, visualization and cooperative operation tools through multiple wavelengths, and realizes the control of optical networks through extended GMPLS protocols and interfaces; - G-lambda project [5] researched by Japan and the United States cooperatively, whose purpose is to achieve the control of optical networks in the Grid Resource Scheduler ( gridresourcescheduler (GRS) and network resource management (network resource management, NRM) systems to establish a standard Web service interface (GNS-WSI) to ensure the collaborative interaction of information between GRS and NRM, and on this basis to realize the establishment of dynamic cross-domain connections and The UCLP (usercontrolled lightpath) program of Canada's CA*net4 research network [6], whose goal is to advocate a "user-enabled network", aims to provide users with the ability to dynamically allocate network resources, and grants them greater ability to innovate network-based applications. the European Union's Phosphorus project [7], which aims at designing and implementing a new web service plane architecture to provide Grid web services for integrated management of network and non-network (computing, storage) resources. Meanwhile, international standardization organizations, such as the Internet Engineering Task Force (IETF), Distributed Management Task Force (DMTF), and Open Grid Forum (OGF), conducted a series of researches on web application and programming environments, architecture, data management, information systems and performance, P2P, scheduling and resource management, and security for Grid Computing.The OGF's Grid High Performance Network ( gridhighperformancenetwork, GHPN) research group has proposed several draft protocols, such as grid-oriented optical network infrastructure (draft-ggf-ghpn-opticalnets-2), networking issues for grid infrastructure (draft-ggf-ghpn-netissues-4), Transport Protocols Overview (draft-ggf-ghpn-transportsurvey-1), Use Cases for Grid Network Services (draft-ggf-ghpn-netservices-usecases) and Grid Network Services (draft-ggf-ghpn-netservices-2), and Grid Optical Burst Switching Networks (draft-ggf-ghpn-GOBS), among others. The Global Light Grid Forum (GLIF) has also recently initiated a series of standardization studies on optical network control plane and grid network interface technologies. In addition, some enterprises (e.g. HP, IBM, Intel, etc.) are also vigorously carrying out research on photonic grid-related technologies and applications (e.g. cloud computing, cloud storage, etc.), and they are investing heavily in setting up data centers all over the world, which all play a more or less important role in promoting the photonic grid technologies and applications. China also attaches great importance to photonic grid technology, the national "863" program, the National Natural Science Foundation of China has set up a number of projects to carry out research on related technologies, and some of the important research technologies include: integrated computing environment for optical networks [8,9], cooperative scheduling of grids and network resources [10,11], photonic grid fault-tolerant techniques, etc [12]. Currently, the common photonic grid architectures mainly include: wavelength grids based on dense wavelength division multiplexing (WDM), dark fiber, and low-cost optical switches; optical burst switching network (OBS)-based grids; and automatically switched optical network (ASON)-based grids. A typical ASON-based photonic grid architecture is shown in Figure 1. The architecture framework is divided into 3 layers. Layer 1 is the application layer, which includes all distributed applications running on the photonic grid. Layer 2 is the service layer, which is the entity of the architecture and includes two parts: workflow and grid middleware. The workflow encapsulates many different application services and publishes them externally. Grid middleware is responsible for downward scheduling and encapsulation of resources, with resource monitoring, resource discovery, resource scheduling, fault tolerance and security control and other functions. Layer 3 is the physical resource layer, which is divided into two parts: one part is the traditional grid information resources, including computing resources, storage resources, display devices, etc.; the other part is the resources specific to the photonic grid, including port resources, node resources, link bandwidth resources, optical resources, etc. The basic workflow is as follows: first, the grid middleware is responsible for downward scheduling and encapsulation of resources, with resource monitoring, resource discovery, fault tolerance and security control. The basic workflow is as follows: first, the service layer obtains the relevant information of the physical resource layer through the relevant interface; when the service layer receives the user request from the application layer, it calls the resource management and scheduling module to assign the tasks of computation, storage, display, etc., to different available resources, and when data transmission is required, it calls the control plane of the optical network to dynamically establish the optical channel connection. Through the above steps, the optimal utilization of resources and the maximum satisfaction of users' QoS requirements can be effectively achieved.5 Application of Photonic Grid TechnologyIt is well known that E-science plays a vital role in the scientific research programs and industrial development in many fields, such as: planetary fluid and magnetohydrodynamic calculations in the field of astronomy; calculations and simulations of a new generation of non-toxic and non-polluting launch vehicles; Numerical wind tunnel and load fatigue calculations in aircraft design; virtual manufacturing and aerodynamic design of the whole vehicle in automobile manufacturing; steel plate collision performance calculations and simulation analysis of steel pipe molding in steel production; thermal and hydraulic analysis of nuclear reactor cores, analysis of nuclear reactor protection and control, and analysis of stress and seismic mechanics of nuclear grade equipment. In these applications, users located in different regions need to *** enjoy data resources, conduct large-scale collaborative computation and analysis, and realize data interaction and transmission of large data streams. A typical example of a photonic grid application is the high-energy physics experiment conducted by CERN, the European Atomic Energy Research Agency, which aims to process petabyte-scale experimental data continuously generated by the Large Particle Collider. The analysis and processing of these data exceeds the capability of any supercomputer or cluster system in the world, so the CERN computer center is responsible for distributing these data to regional centers in Europe, North America, Japan, etc. through high-speed networks, and the latter then further breaks down the task to the desktops of physicists, and completes the relevant experiments through the computation and collaborative analysis of physicists in different regions*** together. The latter then further breaks down the task to the desktops of physicists, and through the computation and collaborative analysis of physicists in different regions, the relevant experiments are accomplished *** together. At present, there are nearly 10,000 scientists located in more than 60 countries and regions in the world to participate in the experiment, the different regions are used between the 10Gbit/s optical network channel for data interaction and transmission. Another application example is the application of real-time very long baseline interferometry (e-VLBI). e-VLBI is a radio interferometry technique that uses a network to transmit observation data from astronomical telescopes to a data processing center in real time for processing. It has important scientific significance and practical value in the fields of spacecraft precision tracking, spaceflight measurement and control, precision time comparison, deep space observation, artificial Earth satellites, lunar probes, and interplanetary probes of the solar system. In the next-generation e-VLBI system, the sampling rate of the radio telescope at the observatory will reach 10 Gbit/s, and the data aggregation rate at the data processing center will reach 40 Gbit/s, and the data need to be transmitted in real time from the observatory located in remote areas to the data processing center through ultra-long-distance, high-speed optical network for relevant processing. In the face of the above application needs, scientists in Europe, the United States, Japan, South Korea, Australia and other countries are carrying out a series of e-VLBI technology research based on high-speed optical networks, such as the European EXPRES research program and the East Asia e-VLBI research program. The NSF-funded GRAGON research program has also conducted relevant research and field experiments for e-VLBI applications regarding the dynamic establishment of optical pathways, large file data transmission, etc. [13]. The photonic grid can break through the network bottleneck of E-science applications, making the wide application of high-performance computing a reality, and the data exchange and information interaction between users and users, and between users and high-performance computers can be realized conveniently and in real time, which will accelerate the user's scientific research process, promote the development of related industries, and bring a bright future to researchers and owners of high-performance computing resources. Photonic Grid can be used to manage valuable instruments distributed all over the world, and by providing means of remote access to the instruments and equipment, it can improve the utilization rate of the instruments and greatly facilitate the use of users. It can also be used to construct networked virtual reality environments to visualize the results of high-performance computations or databases, enabling distributed users to work together in the same virtual space. The environment can be widely used in many fields such as interactive scientific visualization, medical treatment, education, training, art, entertainment, industrial design, information visualization, etc., such as telemedicine, distance learning, virtual history museums, and collaborative learning environments. From the above analysis, it can be seen that photonic grid has a broad application prospect. Photonic grid represents a direction of the development of optical transmission network, reflecting a general trend of network and business application convergence. The study of photonic grid technology and application system will help promote the development of grid applications and the progress of optical network technology. It can be predicted that the photonic grid has a very important theoretical research value and social significance, at the same time has a broad market application prospects, in the economic construction and social development will play an extremely important role.