Topic: Grid
Explanation about the difficulty of topics:
★ Learn about concepts, basic topics
★★ Learning concept, there is scheduled
★★★ Reference example, open problem (answer may not be unique)
Number of topics for difficulty:
★ 17 questions
★★ 22 questions
★★★ 11 questions
topic
1. Explain the concept of five-layer hourglass structure. ★
answer:
The five-layer hourglass structure is an early abstract hierarchy, focusing on the "agreement", emphasizing the status of the agreement in the grid resource sharing and interoperability. A mechanism is implemented through an agreement such that the user and resources of the virtual organization can be used for resource usage, establish a sharing relationship, and further manage and develop new shared relationships. This standardized open structure is very beneficial to the extensibility of the grid, interoperability, consistency, and code sharing. The figure below shows a typical structural diagram of a five-layer hourglass structure.
The five-layer structure is shaped as an hourglass, which is caused by uneven distribution of the number of various agreements. Considering the core transplant, the convenience of the upgrade, the number of protocols in the core part is relatively small (for example, TCP and HTTP on the Internet), for its core part, to implement the upper layer protocol (top floor of the hourglass) to the core protocol At the same time, the map of the core protocol to the next protocol (the bottom of the hourglass) is realized. According to the definition, the number of core protocols cannot be too much, so that the core protocol has become a bottleneck of protocol hierarchy. In the five-layer structure, the resource layer and the connecting layer together constitute this core bottleneck portion, which promotes separate resource sharing.
2, please explain the concept of OGSA. ★
answer:
Open Grid Services Architecture OGSA (Open Grid Services Architecture) is an important standard recommendation for Global Grid Forum 4, which is the most important after the five-layer hourglass structure, and the latest grid architecture, which is the next generation. structure. Among them, our mind is: a service-centric model; the same web service framework; breakthrough technology applications. Two major support technologies for building OGSA are grid technology (such as Globus software packages) and Web Service. Five-layer models are equivalent to the interoperability, but OGSA emphasizes the service's view, transforming interoperability issues to define services interface and identify protocols that activate specific interfaces.
3, the five-layer hourglass structure is () from bottom to (). ★
(A) Construction layer, resource layer, connection layer, aggregation layer, application layer
(B) Construction layer, connection layer, resource layer, aggregation layer, application layer
(C) Construction layer, connection layer, aggregation layer, resource layer, application layer
(D) resource layer, connection layer, structure layer, aggregation layer, application layer
(E) Application layer, connection layer, aggregation layer, structure layer, resource layer
answer:
Choose B.
4. The following describes the description of each layer of the five-layer hourglass structure is (). ★
(A) Construction layer: Define some of the single resource sharing operation protocol
(B) Continental layer: Define the communication and verification protocols for GRID network transactions, based on Internet protocol
(C) Resource layer: Provide a set of tools and interfaces for local resource control
(D) Polymerization: Provide multiple resource collaborative work
(E) Application layer: Provide GRID system development and application development tools, environment
answer:
Choose AC.
5. What is the three standards that IAN FOSTER proposed? ★ A:
Grid must meet three conditions:
(1) Collaborate use resources in an environment of non-centralized control;
(2) Use standard, open and universal protocols and interfaces
(3) Provide non-flat service.
(Refer to July 2002 Ian Foster "What is Grid? Judgment Whether Grid's Three Standards")
6. In the grid service interface listed below, () is an existing service interface, where () the interface is necessary. ★
(A) GridService
(B) Registry
(C) Factory
(D) WebService
(E) HandleMap
answer:
Select (ABCE), select (a).
7. What are the grid core services? ★
answer:
(1) High performance scheduling technology;
(2) High throughput rate resource management technology;
(3) Performance data collection, analysis and visualization techniques;
(4) Safety technology.
8. Grid applications are mainly divided into () these major parts. ★
(A) Distributed supercard
(B) Distributed instrument system
(C) distributed storage system
(D) remote immersion
(E) Data-intensive calculation
answer:
Choose ABDE.
9. What is the implementation of the GLOBUS project? What are the main components of the GLOBUS system? ★
answer:
The Globus project was launched in the mid-1990s, the original purpose is to connect the various high-performance computing centers in the United States to connect through high-performance networks, so that the US universities and research institutions use, improve the efficiency of high performance computers. At present, the Globus project integrates the Web Service technology in the field of commercial computing. It is intended to be limited to the scientific calculation field, but also a wide range of commercial applications, basic network environment support, more convenient information. Sharing and interoperability, thereby deeply affecting business models, personnel working methods and lifestyle. Globus's current main job is to establish a general protocol that supports grid computing, develops services that support grid computing services, and implement software development tools for support grid computing environments.
The main components of the Globus system are:
(1) Grid Safety Infrastructure (GSI), responsible for secure authentication and encrypted communication under wide area network, providing single sign-on function, remote identification, data transfer encryption, etc., providing a GSI protocol interface, It is a core part that guarantees mesh computing security.
(2) Globus Resource Assignment Management (Gram), responsible for resource request processing, remote task scheduling, remote task management, etc., responsible for resource description language information, is the task in the grid computing environment Execution Center.
(3) Metal Calibration Directory Service (MDS), mainly completes the discovery, registration, query, modification of information in the grid computing environment, providing a real, real-time dynamic response to the grid computing environment, is grid calculation Information Service Center in the environment. (4) Global Level 2 Storage Service (GASS), simplifies the operation of the application for remote file I / O in the Globus environment, making it basically don't have to change the application of UNIX and standard C language I / O libraries. ZTE is a middleware that supports grid computing environment remote I / O access.
(5) Grid FTP service is a high-performance, secure, reliable data transfer protocol, and optimized the high bandwidth wide network environment, which is a data transfer tool in the grid computing environment.
(6) Globus replication management, which can only place some relevant data in the nearby position of the scientific application, so that the scientific application can quickly access the data.
10, what is the goal of grid, grid? ★
answer:
The grid is a concept that is borrowed from the power grid. It was originally desirable to calculate and compute resources. "Open the power switch can be used", do not need to care for who, how to provide these services. The goal of the grid is abstract and quantifies the computing resource, and it is possible to complete the "quantification" calculation through the network "quota", anytime, anywhere.
11. What is the essential characteristics of the grid? ★
answer:
The essential characteristics of the grid are: distribution and resource sharing: distribution is the most characteristic of the grid, the grid is to complete the calculation by centralized resource, and the sharing of resources is a means of concentrated resources. Height abstract: Put calculation And all computational resources are highly abstract to become the "power wiring board" visible to the user, and the other things are transparent to the user. Self-similarity: There is the same or similar law dynamics and diversity in large scale and small scale: Like the power grid, the user's demand is changed, so the dynamicity is a basic problem autonomousness that the grid needs to be considered. Multipleness of management: Autonomous and external controlled integration inside the grid node is a feature of the grid, and the hierarchical resources require hierarchical management, and the hierarchical considerations from the homepacted nodes. .
12. The advantages of grids reflected in the application are mainly,. ★
answer:
High performance, low cost.
13. The main architecture of the grid? ★
answer:
Foster, etc., the five-layer hourglass, which is proposed earlier, starting from the bottom layer as a structural layer, a connection layer, a resource layer, a gathering layer, and an application layer, which is an early mesh architecture, and later OGSA's emergence makes this system. The structure has been dimmed. The grid architecture model proposed by the US Globus project adopts local service layer, core service layer, high-level service and tool layer, and landscaping. On this basis, the United States's Argonne National Laboratory, the University of Chicago, the University of South California and IBM proposed an open grid service architecture (OGSA). OGSA uses a fiber layer, a contact layer, a resource layer, a collaborative layer, and a layer of five-layer structure.
14. What is the origin of Grid? ★
answer:
The word "Grid" comes from the Power Grid of people. The current demand for calculation is continuously growing. High-speed network, function and performance is more powerful, large-scale computing system, high-performance storage system, and advanced data analysis software to solve the problem of calculation in business areas and scientific fields. But using these new capabilities means unprecedented challenges brought about by increasing workload requirements. When an organization is working hard to maintain competitiveness, they face many unprecedented challenges: reduce computing costs, increase computational throughput, and speed up the response to the market, improve production service quality and product innovation are very important. An organization's investment in hardware needs to be careful, and you must find a way to use available resources to complete more work. Although the demand for calculation is very large, many existing systems are not fully utilized. Some individual servers may be full of load, but most systems are not full-load work, so that many computing cycles are not used, but in fact these computing cycles can be used to meet growing computational resource requirements. Grid computing allows organizations to more efficiently and flexibly with their distributed computing resources, from existing systems to help organizations gain competitive advantage. 15. Globus project grid security infrastructure GSI is important in the network and emphasizes the integration with existing distributed security technology. GSI is based on an encryption system, adopting certification and communication protocols, and has expanded them, allowing GSI to support single sign-on. ★
answer:
Transport layer, application layer, public key, X.509, SSL.
16. In the following options, which are common grid applications development tools (). ★
(A) Cog Kits
(B) gt3
(C) Java Cog Kit
(D) MPICH-G2
(E) CORBA
answer:
Select ABCD.
17. What are the main components of Globus's data management? ★
answer:
(1) Global secondary storage service GASS, through GASS can simplify the operation of the application to the remote file I / O in the Globus environment, so that the application using UNIX and standard C language I / O libraries can basically do not change. Operating in the Globus environment;
(2) Grid FTP service, GridftP supports third-party transmission, breakpoint resume, parallel transmission, combined with GSI, caching, etc., is a data transmission tool in a grid computing environment;
(3) Globus replication management, replication management is an important issue that needs to be considered in a wide range of scientific applications. Due to the existence of access to large remote files, the Globus replication directory is closest to the scientific application by putting some relevant data smart. Location, the scientific application can quickly access the data quickly.
18. The first database application server for grid computing has officially facing the city, this product is (). ★★
(A) Microsoft SQL 2004
(B) Oracle 10g
(C) GridBase
(D) db2 for grid
(E) Sybase
19. Five major applications of grid: distributed supercarcies, distributed instrument systems, data intensive calculations, remote immersion and information integration. In fact, these five applications can also be implemented on the current World Wide Website, where is the advantage of the grid in these applications? ★★ A:
The advantage of the grid is reflected in the request of each item. The grid on demand represents a quantified basic unit, formatting resource sharing, formal public protocol, and axiological combination innovation. Grid is visible, spectrum, expansion, and controlled strategic innovation tools that can promote smart computers and networks from the architecture.
20. What are the basic requirements of the grid architecture to include? ★★
answer:
(1) Provide a single distribution of resources for the application and aggregated various distribution of autonomous resources into a system, cannot destroy the existing system, and do not cause the application's rewriting;
(2) Resource sharing interoperability between different organizations is essential, you can directly access resources without just file sharing and exchange, which is embodied in various resource components and mechanisms;
(3) There is a standard protocol to be critical. Under the premise of compliance with the standard protocol, you can allow the cross-node to work, allow the underlying organizational sharing to become possible; there is no standard protocol, interoperability and sharing difficult Implementation, repetitive development and installation inevitable;
(4) From the point of view of the Grid implementation, the grid architecture is a set of hierarchical middleware services, according to the design principles of modern software systems, providing support for application tools and applications in the way of service, while serving Implementation must be able to use all levels of functions and services to be constructed.
21. The following figure is an example of a data mining, which is based on the structure, and the authentication is done in the first step in the illustration. ★★
answer:
OGSA structure, 2.
22. Interpret the grid calculation and the Web Service relationship. ★★
answer:
Grid calculation is closely related to Web Service: The core of Web Service is to connect various applications on a large heterogeneous network, with the WEB Standard UDDI, WSDL, and XML / SOAP, etc. The network has further developed to an application platform. When more and more Web Service implementations, the various features of the application will change. Some applications can be dynamically constructed from availability as needed or based on the functions available.
OGSA has absorbed many Web service standards, such as Web Service Description Language (WSDL), Simple Object Access Protocol (SOAP), Light Directory Access Protocol (LDAP), Web Service Probe (WS-Inspection). These standards are used to position, scheduling computing resources and ensuring their security. The advantages of OGSA use the Web service framework are: On the one hand, through registration and discovery interface definitions and terminal implementation descriptions and dynamically generate specific interface binding agents, it is possible to support service dynamic discovery and synthesis in heterogeneous environments. On the other hand, since the Web service mechanism is widely used in commercial applications, the OGSA uses a Web service framework to enable people to use many existing development tools and extended services, such as WSDL processors that generate language binding.
Some people think that Web Service and Grid Service need to discover services, and then bind the service to integrate into the app, so that the Web Service and Grid Service are one thing, this is wrong. Web Service emphasizes the discovery and calls of permanent service, Web Service's service is stateless, but in the grid environment, you must also support instantaneous service instances, dynamically create and cancel. It can be said that Grid Service just uses a layer of outer skin of Web Service. 23. The picture below is a system or OGSA system of a five-layer hourglass and says your reason. ★★
answer:
Osga structure. Since it can be seen from the figure that this system implements a must-have GRID Service interface, and the collaboration of Interface Provider by Miner Factory, Database Factory can support the creation of Linxi services.
24. OGSA has four main levels: physical and logical resources, define grid services Web services and OGSI extensions, based on OGSA architecture, and grid applications. How do you understand this sentence? ★★
answer:
Not exactly correct. This sentence should be a service grid, but it is not necessarily suitable for other grids, such as compute grids.
25. What is the difference between grid and distributed system? ★★
answer:
Grid computing is a new high-performance distributed computing method. Grid computing as a new generation of distributed computing methods, the main difference with traditional distributed calculations is that by large-scale sharing of computing resources without centralized control mechanism, meet the application for high performance computing requirements, and this A large-scale sharing of computing resources is dynamic, flexible, safe and collaborative.
26. The purpose of scheduling is to make a large number of applications to achieve maximum performance while sharing grids. Where is the dispatching strategy and technology in grid scheduling relative to traditional high performance calculations? ★★
answer:
Grid scheduling is more complicated relative to scheduling techniques and strategies in traditional high performance calculations. This is because any grid scheduler cannot manage all grid resources, but can only make grid resources within a certain range. In addition, grid resources are still dynamic. Since other applications have a large impact on performance, and often, the complex diversity of grid resources, different types of resources will show different performance characteristics, and the same type of resources are displayed by sharing. It is also changed over time. Therefore, the scheduling of the grid requires the establishment of performance prediction models that make full use of the dynamic information of the grid to represent the fluctuation of the mesh performance, and the mesh scheduling must take into account a variety of environmental and conditions, this is It is determined by the characteristics such as mesh isomer and diversity. In grid dispatch, there is also problems such as transplantability, scalability, efficiency, repeatability, and grid scheduling and local scheduling combination, which are issues that need to be solved with grid technology.
27, explain your understanding of the same and differences between clusters and grids. ★★
answer:
They are all distributed systems, many places, sharing, and collaboration, and similar places in technology. However, the conditions and environment are different, the cluster is generally related (also isomeric), and the region is very close, the stronger the performance, the closer the performance. Typically in a room, at least within a local area network. The grid is much wide. Regardless of the breadth, or the complexity of the system. The composition of the grid can be complicated, including dedicated memory, various instrumentation, etc., which also include clusters, various types of hardware, a variety of different system software, and complex middleware. From a distributed system, the most concentrated way is that the cluster, the most scattered method is the grid. It belongs to both ends of the distributed. 28. How to classify the grid and explain it separately. ★★
answer:
(1) Computing grid: emphasize the consolidation of calculation, management, etc.
(2) Data grid: emphasize data storage, management, transmission, and processing grid;
(3) Access grid: emphasize information storage, management, transmission, and processing grids;
(4) Information grid: emphasize the grid of data storage;
(5) Service Grid: emphasize the grid integration of application services;
(6) Knowledge Grid: emphasize knowledge storage, management, transmission, and processing grids;
(7) Semantic Grid: emphasizes the mesh of semantic parsing, realizing semantic interoperability;
(8) Access Grid: Provide a single access channel to get the grid of the grid resource;
(9) Grid portal: Provide a single access channel.
29, introduce some abroad of major grid projects? ★★
answer:
The goal of the American scientific network project is to establish a scientific computing infrastructure (ADCI) based on grid middleware and system tools to enable energy-based scientific computing systems to meet the internal scientific research task requirements in the energy department. The goal of remote distributed computing and communication projects is to create a workable grid for accessing the three weapons labs of the US Department of Energy to support remote computing and distributed calculations in defense programs. Analysis, design, manufacturing, certification.
Earth System Grid II, ESG project is jointly borne by scientists from five national laboratories such as Argonne National Laboratory. The main goal is to address the huge challenges faced by global earth system model analysis and discovery knowledge, providing a seamless powerful virtual synergistic environment for next-generation climate research.
The goal of the National Earthquake Project Simulation Grid enables seismologists to conduct a large-scale computing task of various seismic engineering experiment data analysis.
In addition, the United States is currently underway computing R & D projects also include Globus, the Information Power Grid of NASA (IPG), National Technology Technology Grid, Virtual Laboratory Virtual Laboratory Project, an Astrophysics Simulation Collaboratory, ASC, International Virtual Data Grid Laboratory, IVDGL, and more. Globus goals build a grid software infrastructure. The research content includes resource management, data access, application development, grid security, etc. Many users in the world use Globus Toolkit to create grid and develop grid applications.
The goal of the information power grid is to provide continuous and reliable calculation power sources for the scientific research tasks of NASA.
The goal of the US National Technical Network Project is to create a seamless integration of collaborative computational environment prototype systems.
Virtual Laboratory project is committed to research, design, and develops techniques that help solve data dense, involving large-scale computable molecular biological problems.
The main goal of the celestial physical simulation cooperation laboratory project uses the research results of CACTUS and GLOBUS grid to provide large-scale parallel computing power for advanced visual interaction and metamorphism, and achieve large-scale simulation in the field of heaven-body physics.
The International Virtual Data Network Laboratory project is created by the European Data Grid, the United States's grid physics network, and the Particle Physics Data Grid.
30, what is P2P calculation, what is the relationship with the grid? ★★
answer:
Peer (P2P) calculation is a new mode for implementing distributed computing on the Internet. In this mode, the server and client boundaries disappear, all nodes on the network can share computing resources for other nodes "equality". IBM is the following definition: peer-to-peer collaboration, and at least one of the following features: The system is actively collaborated by the marginalization (non-central server) device, each member directly from other Members are benefited from the participation of the server; members of the system simultaneously play the role of the server and the client; the user of the system application can be aware of each other, constitute a virtual or actual population. The grid uses the P2P calculation architecture.
31, what is the relationship with the grid? ★★
answer:
The metad computing technology is a set of computing resources that are different from the nature of the wide area network, and provides a computing service to the user as a separate computing environment. The goal of the metadata system is to enable the user to share the computational resources and calculate in cooperation. The front end of the yuan mainly solves the problem of the end users to use the various types of computing resources on the wide area online by unified interface; the yuan computing kernel mainly solves the problem of collaboration completion on various supercomputers in the wide area network, providing a complete program development and running surroundings. When the user proposes a computing request, the execution code of the calculation problem is reasonably scheduling and managing on the computing resources within the system, and finally the operation results are obtained, and the front end is given to the end user. A good metamorphism system is mainly composed of three parts: one is as simple as possible and reliable use interface; the other is the resource management system; the third is a good programming model and a highly efficient runtime environment. Among them, the resource management system is in the core, it is responsible for the description, change, allocation and protection of various resources across the system, and work closely with the runtime system to ensure the normal operation of the entire system. The resource management system needs to provide strong tools or interfaces for different levels of system administrators to control and protect them. Yuan calculation is the primary form of grid calculation. 32. What is the relationship between the grid and .NET, CORBA, JINI, JAVA RMI, J2EE? ★★
answer:
Grid fuses the RMI, J2EE thoughts of .NET, CORBA, JINI and Java, and extended from software to hardware such as CPU, storage devices, instruments.
33. What is the relationship between OGSA and OGSI ★★
answer:
The OGSI specifies the interface used when sending a processing request to the grid, which is equivalent to WSDL in the web service, in the core status in grid computing related standards. The open grid service architecture uses fiber layers, contact layers, resource layers, collaboration layers, application layer five-layer structure, and mesh is constructed by calling services, which involves interface issues.
Ogsa is an abstract thing, there is no specific content, OGSA's launch is early over OGSI, and OGSI is the mainly avatar, standardization of OGSA. Globus is an implementation based on OGSI standards. WSRF is the same level of OGSI, which is an absorbed Web Service's reconstruction and development of OGSI.
34. What is the relationship between grid and middleware? ★★
answer:
Grid Middleware is a middleware applied in the grid; the current Globus Toolkit is a grid middleware that connects the heterogeneous host nodes to the Globus grid.
35. What is the relationship between the grid and IBM proposed independent computing? ★★
answer:
The idea of "autonomous computing" is the increasingly complex IT system management of background, automation, and intelligence, and transforms heterogeneous networks into true harmonious autonomous digital nervous systems through powerful middleware. The grid itself is an autonomous calculation.
36. How do domestic grids have those projects, how is the goal of each project, what is the current progress? ★★
answer:
(1) The goal of the ACI project is to connect distributed high-performance computers, valuable instruments, databases such as different geographic locations, to constitute a virtual computer, and users share resources, discuss, and cooperate to carry out research projects.
(2) Weaving actresses, VEGA programs, based on metadata, component framework, intelligent body, grid public information protocol and grid computing protocols as the main research breakout point.
(3) Shanghai grid, the marine governments will cooperate with IBM companies to establish "urban grid", integrate government information resources in the city, and improve the public service capacity and efficiency of the government.
37. The following figure is a schematic diagram of the acquisition process of the safety certification certificate in the grid environment. Please mark the procedure of the certificate acquisition process in the circle in the figure. ★★ A:
38. The picture below is a schematic diagram of the security authentication process in the Globus environment, and uses the number to label the order of the process, please add a description of each step. ★★
answer:
39. What is the main study of the safety of mesh? ★★
answer:
(1) Integration Program: Focus on how to make full use of existing services and interfaces and abstracts to be extended;
(2) Collaborative Working Ability Solution: Focus on how to call problems in the virtual organizations with different security mechanisms and strategies;
(3) Trust strategic program: Focus on how to define, manage, and execute trust policies in dynamic grid environments.
40. Schematic of the GSI certification process of the following picture below. The figure shows the process of host B authentication host A, but the authentication process step order in the circle in the schematic diagram is being applied, please fill in the correct step order. ★★★
answer:
41. What are you thinking about if the GSI authorization still exists? ★★★
answer:
(1) GSI requires the global user of each access resource to have a book on the local resource server, each resource server needs to maintain a huge clumsy global / local mapping table, which is difficult to extend to ownership A large-scale environment of a large number of resources and a large number of users.
(2) GSI lacks access control mechanisms with good scalability based on global strategies.
(3) For issues in the current authorization mechanism in the current large-scale grid environment, we propose community authorization strategies in virtual organizations.
42, talk about your suggestion of building a grid. ★★★
answer:
(1) Strive for national and government support.
(2) Strive for the support of big companies.
(3) Creating business opportunities and utilizing market and economic levers.
(4) From part to overall.
(5) The same agreement, specification.
43. Some people propose the grid as a computing power, whether a metering unit can be proposed as a metrics used. Like the power network, the user has a separate electric meter, paying according to the amount of electricity. Talk about this view of this point. ★★★
answer:
(Answer a) should pay. Good service grids can charge appropriate fees. Such as: The computing grid can be paid according to the calculation capabilities used, and the storage grid can charge the fee according to the storage size and service time.
(Answer b) should not be paid. The grid name is the next generation of the Internet, you can refer to the current Internet to determine if the grid can be used. On the Internet, people provide significant resources and services are resources. Others use these resources and services that do not pay, so the grid should not pay.
44. Will it be the most infrastructure of the network in N-year? Give your opinion. ★★★
answer:
A answer: Will not, just like the Internet does not replace the PC. They are at different levels to solve different problems. Grid is based on web, inevitably built on the Internet.
45, the grid is becoming the next generation of the Internet. Please refer to Grid Technology Analysis, the computer as a grid device is improved in terms of the computer under the grid era. ★★★
answer:
1. Safety: Compared to the network environment, the computer is more in the grid, which has a higher security performance requirement for a single computer that appears in the face of grid devices. 2, individual and collective relationship: Each computer (grid device) is a member of the grid collective. For mesh applications, the grid is a whole, regardless of the task (or application), is all submitted, all Executed with resources of the entire grid. That is, a task running on a large computer can run through or even a mobile phone. In the grid, the individual features will be able to minimize in the grid society. That is, we can run an application on the PC computer, such as Word, you can run the same application on your phone, and even run half of Word on the PC, transfer to the mobile phone, and the grid application itself There is no spot change.
Therefore, in order to adapt to the grid environment, the current computer needs to improve security and facilitate distributed calculations, and it is necessary to improve from hardware, operating system, application architecture, communication and other aspects.
46. A unit requires the previous system, demand is: 1. Most of the systems that need to be integrated are database systems; 2. The ability to respond real-time response to information; 3, the organization's organization requires grading. Consider whether the unit is suitable for using a grid platform to solve this problem? ★★★
answer:
The grid is not good enough to support the database. It is not very appropriate to develop Grid Service. It does not have universality; most of the grid is still a passive system, that is, the user is required to apply for a service, it is one The way of dragging, and more demand for this unit needs to be a way to pull; and the model used in the grid is more models, not a grading model.
The issue of this unit should have ready-made commercial software to complete, but the price should be more expensive. There is no ready-made grid platform to simply integrate the unit's system into and meet the requirements. If you want to do it yourself, take into account development costs, maintenance, etc., the price is bound to be higher, so use the grid platform is not suitable. of.
47. If the grid is compared to a virtual supercomputer, the grid calculates the function of the CPU, the data grid is similar to the function of the disk, then what is the data line, motherboard, and graphics card? ★★★
answer:
There is a misunderstanding here: We can't see the mesh as a middleware when it is a grid computer. When the grid is viewed as a middleware, the grid cannot be a computer because the original computing model has not changed.
The grid computer is a virtual computer built on the "distributed" system, so its computational model must have a fundamental change, ie, is not a Feng Dangman agency, nor should it be a map forever model.
If you must match the concept of "data cable, motherboard, graphics card" in the grid environment to the specific grid device, you can understand: "Data Line, Motherboard" is a communication network and service registration station, "graphics card, Display is a PC terminal. However, this match is a bit thrown.
48. Grid is a new technology, any new technology or less ingredients with foam? Are you agreeable? Does grid technology have a foam color? In general, the current development of mesh technology is what is in one situation? How far is grid from practical applications, how far is it from commercial applications? Grid needs to go through several phases to go to the real commercial? Grid wants to get a wide range of business applications, will you encounter some obstruction? ★★★
answer:
(1) Do not agree, the technology itself does not have a foam, and the foam is commercial speculation. It has already begun commercial hype. (2) Technically, in the research and development stage, business is in the preheat period.
(3) 2 to 3 years.
(4) Prototype application system development, grid standardization (normative, agreement), mature business model.
(5) Yes, such as security issues, billing issues, etc.
49. The universal form of computing grid in the world is that the task can be distributed across the nodes, but the current phase network conditions and research levels can not effectively support this idea, so some people can make use of grid " Calculate pool. "Grid Calculation Pool" has the following features: (1) Calculating resource sharing; (2) Do not break up a task into n sub tasks, but just schedule one of one of the appropriate machines; (3) Submit tasks through the Web And view the results. Please give your evaluation from the three characteristics of "Grid Calculation Pool". ★★★
answer:
(1) Computing resource sharing can improve the utilization and quality of service, but it is necessary to prove;
(2) There is only a very little task that any single high-performance computer cannot be processed. This definition can avoid the complexity of the decomposition task (while reducing the threshold of the user using high performance calculations), and more importantly, the performance of frequent communication between the task submodules is avoided.
(3) Submit tasks and view results via Web. This can greatly facilitate the user to access high-performance computers: users don't have to come, do not have to queue in the laboratory, and can be submitted and running and running at home. Moreover, in general, the internal traffic of the task submits and the return result is small compared to the calculated amount of traffic.
50. Depending on your understanding of the storage mesh, design a virtualized storage system within the local area network. ★★★
answer:
The system structure is shown in the figure.
The entire system mainly includes storage service point SSP (Storage Service Point), global naming server GNS (Global Name Server), Explorer RM (Resource Manager), Certification Center CA (CERTIFICATE Author), client, storage agent SA (Storage Agent ) And visual management.
SSP is the entrance to the entire system, all access to the system through SSP, which mainly provides an FTP interface, a CA interface, an RM interface, and a GNS interface; the number of SSPs in the system can be increased as needed; SSP take over the traditional solution Partial features of GNS reducing GNS loads, enhances system scalability.
GNS is responsible for the metadata management of the system, mainly including metadata operation interface, metadata fault tolerant system, metadata search system.
RM includes resource scheduling modules and copy management modules, which are primarily responsible for the application and scheduling of resources while providing transparent copies creation and selection policies.
The client currently supports three forms: universal FTP client, file access interface, and special client. The user can also achieve more performance services, but also capable of performing user group operations, not only capable of user group operation, search and sharing, etc., but also available services.
The CA includes a certificate management system, which is mainly responsible for the security and data access control of the system, and it records the user's registration information.
SA blocks the diversity of storage resources, providing a unified storage access interface for the system while providing a file operation mode and an extended FTP operation method, and it provides support for file replication management operations to provide high-efficiency transmission.
