Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.507907
O. Damani, V. Garg
We propose a new algorithm for recovering asynchronously from failures in a distributed computation. Our algorithm is based on two novel concepts-a fault-tolerant vector clock to maintain causality information in spite of failures, and a history mechanism to detect orphan states and obsolete messages. These two mechanisms together with checkpointing and message-logging are used to restore the system to a consistent state after a failure of one or more processes. Our algorithm is completely asynchronous. It handles multiple failures, does not assume any message ordering, causes the minimum amount of rollback and restores the maximum recoverable state with low overhead. Earlier optimistic protocols lack one or more of the above properties.
{"title":"How to recover efficiently and asynchronously when optimism fails","authors":"O. Damani, V. Garg","doi":"10.1109/ICDCS.1996.507907","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.507907","url":null,"abstract":"We propose a new algorithm for recovering asynchronously from failures in a distributed computation. Our algorithm is based on two novel concepts-a fault-tolerant vector clock to maintain causality information in spite of failures, and a history mechanism to detect orphan states and obsolete messages. These two mechanisms together with checkpointing and message-logging are used to restore the system to a consistent state after a failure of one or more processes. Our algorithm is completely asynchronous. It handles multiple failures, does not assume any message ordering, causes the minimum amount of rollback and restores the maximum recoverable state with low overhead. Earlier optimistic protocols lack one or more of the above properties.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128938993","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.507914
Jie Wu
Reliable communication in cube-based multicomputers (including disconnected ones) using the safety vector concept is studied in this paper. In the proposed approach each node in a cube-based multicomputer of dimension n is associated with a safety vector of n binary numbers which is an approximated measure of the number and distribution of faults in the neighborhood. The safety vector of each node in an n-dimensional hypercube can be easily calculated through n-1 rounds of information exchange among neighboring nodes. Optimal unicasting between two nodes is guaranteed if the kth bit of the safety vector of the source node is one, where k is the Hamming distance between the source and the destination. An extended deadlock-free unicasting using virtual channels is also introduced.
{"title":"Reliable communication in cube-based multicomputers using safety vectors","authors":"Jie Wu","doi":"10.1109/ICDCS.1996.507914","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.507914","url":null,"abstract":"Reliable communication in cube-based multicomputers (including disconnected ones) using the safety vector concept is studied in this paper. In the proposed approach each node in a cube-based multicomputer of dimension n is associated with a safety vector of n binary numbers which is an approximated measure of the number and distribution of faults in the neighborhood. The safety vector of each node in an n-dimensional hypercube can be easily calculated through n-1 rounds of information exchange among neighboring nodes. Optimal unicasting between two nodes is guaranteed if the kth bit of the safety vector of the source node is one, where k is the Hamming distance between the source and the destination. An extended deadlock-free unicasting using virtual channels is also introduced.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128618886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.508013
Sajal K. Das, M. C. Pinotti, F. Sarkar
We efficiently map a priority queue on the hypercube architecture in a load balanced manner, with no additional communication overhead. Two implementations for insert and deletemin operations are proposed on the single-port hypercube model. In a b-bandwidth, n-item priority queue in which every node contains b items in sorted order, the first implementation achieves optimal speed-up of O[min{log n, b(log n)/(log b+log log n)}] for inserting b pre-sorted items or deleting b smallest items, where b=O(n/sup 1/c/) with c>1. In particular, single insertion and deletion operations are cost-optimal and require O(log n/p+log p) time using O(log n/log log n) processors. The second implementation is more scalable since it uses a larger number of processors, and attains a 'nearly' optimal speed-up on the single-port hypercube. The insertion of log n pre-sorted items or the deletion of log n smallest items requires O(log log n)/sup 2/ time and O(log/sup 2/ n/log log n) processors. However, on the slightly more powerful pipelined hypercube model, we are able to reduce the time complexity to O(log log n) thus attaining optimal speed-up. To the best of our knowledge, our algorithms provide the first implementations of b-bandwidth distributed priority queues, which are load balanced and yet guarantee optimal speed-up.
{"title":"Distributed priority queues on hypercube architectures","authors":"Sajal K. Das, M. C. Pinotti, F. Sarkar","doi":"10.1109/ICDCS.1996.508013","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.508013","url":null,"abstract":"We efficiently map a priority queue on the hypercube architecture in a load balanced manner, with no additional communication overhead. Two implementations for insert and deletemin operations are proposed on the single-port hypercube model. In a b-bandwidth, n-item priority queue in which every node contains b items in sorted order, the first implementation achieves optimal speed-up of O[min{log n, b(log n)/(log b+log log n)}] for inserting b pre-sorted items or deleting b smallest items, where b=O(n/sup 1/c/) with c>1. In particular, single insertion and deletion operations are cost-optimal and require O(log n/p+log p) time using O(log n/log log n) processors. The second implementation is more scalable since it uses a larger number of processors, and attains a 'nearly' optimal speed-up on the single-port hypercube. The insertion of log n pre-sorted items or the deletion of log n smallest items requires O(log log n)/sup 2/ time and O(log/sup 2/ n/log log n) processors. However, on the slightly more powerful pipelined hypercube model, we are able to reduce the time complexity to O(log log n) thus attaining optimal speed-up. To the best of our knowledge, our algorithms provide the first implementations of b-bandwidth distributed priority queues, which are load balanced and yet guarantee optimal speed-up.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114459095","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.507916
N. Islam, Amitabh Dave, R. Campbell
Parallel programs running on top of generic protocols (e.g. TCP) in a cluster of workstations often do not perform or scale as well as one would expect. One reason for this is that both the performance and scalability of parallel applications are highly dependent on the speed of communication, yet the generic protocols used to guarantee reliable message delivery add unnecessary overhead which degrades the performance of the parallel application. The main thesis we explore in this paper is that it is possible to use knowledge of application behavior to design protocols that are more efficient. In particular, we investigate automatic techniques for generating optimized application-specific network protocols for parallel applications running on unreliable networks. Our algorithms assume that the application communication can be represented by a context free grammar. Such algorithms form the basis for a communication compiler.
{"title":"Communication compilation for unreliable networks","authors":"N. Islam, Amitabh Dave, R. Campbell","doi":"10.1109/ICDCS.1996.507916","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.507916","url":null,"abstract":"Parallel programs running on top of generic protocols (e.g. TCP) in a cluster of workstations often do not perform or scale as well as one would expect. One reason for this is that both the performance and scalability of parallel applications are highly dependent on the speed of communication, yet the generic protocols used to guarantee reliable message delivery add unnecessary overhead which degrades the performance of the parallel application. The main thesis we explore in this paper is that it is possible to use knowledge of application behavior to design protocols that are more efficient. In particular, we investigate automatic techniques for generating optimized application-specific network protocols for parallel applications running on unreliable networks. Our algorithms assume that the application communication can be represented by a context free grammar. Such algorithms form the basis for a communication compiler.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124273884","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.507898
R. Grimm, Wilson C. Hsieh, W. D. Jonge, M. Kaashoek
Atomic recovery units (ARUs) are a mechanism that allows several logical disk operations to be executed as a single atomic unit with respect to failures. For example, ARUs can be used during file creation to update several pieces of file meta-data atomically. ARUs simplify systems, as they isolate issues of atomicity within the logical disk system, ARUs are designed as part of the Logical Disk (LD), which provides an interface to disk storage that separates file and disk management by using logical block numbers and block lists. This paper discusses the semantics of concurrent ARUs, as well as the concurrency control they require. A prototype implementation in a log-structured logical disk system is presented and evaluated. The performance evaluation shows that the run-time overhead to support concurrent ARUs is negligible for Read and Write operations, and small but pronounced for file creation (4.0%-7.2%) and deletion (17.9%-20.5%) which mainly manipulate meta-data. The low overhead (when averaged over file creation, writing, reading, and deletion) for concurrent ARUs shows that issues of atomicity can be successfully isolated within the disk system.
{"title":"Atomic recovery units: failure atomicity for logical disks","authors":"R. Grimm, Wilson C. Hsieh, W. D. Jonge, M. Kaashoek","doi":"10.1109/ICDCS.1996.507898","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.507898","url":null,"abstract":"Atomic recovery units (ARUs) are a mechanism that allows several logical disk operations to be executed as a single atomic unit with respect to failures. For example, ARUs can be used during file creation to update several pieces of file meta-data atomically. ARUs simplify systems, as they isolate issues of atomicity within the logical disk system, ARUs are designed as part of the Logical Disk (LD), which provides an interface to disk storage that separates file and disk management by using logical block numbers and block lists. This paper discusses the semantics of concurrent ARUs, as well as the concurrency control they require. A prototype implementation in a log-structured logical disk system is presented and evaluated. The performance evaluation shows that the run-time overhead to support concurrent ARUs is negligible for Read and Write operations, and small but pronounced for file creation (4.0%-7.2%) and deletion (17.9%-20.5%) which mainly manipulate meta-data. The low overhead (when averaged over file creation, writing, reading, and deletion) for concurrent ARUs shows that issues of atomicity can be successfully isolated within the disk system.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"133 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125057406","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.507968
J. Cobb, M. Gouda
We present a group routing protocol for a network of processes. The task of the protocol is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring each group member is included in the tree. To build this tree, the group routing protocol relies upon the unicast routing tables of each process. Thus, group routing is a composition of a unicast routing protocol, whose detailed behavior is unknown but its basic properties are given, and a protocol that builds a group tree based upon the unicast routing tables. The design of the group routing protocol is presented in three steps. First, a basic group routing protocol is presented and proven correct. Then, the protocol is refined twice, strengthening its properties with each refinement. The final protocol has the property of adapting the group tree to changes in the unicast routing tables without compromising the integrity of the group tree, even in the presence of unicast routing loops.
{"title":"Group routing without group routing tables","authors":"J. Cobb, M. Gouda","doi":"10.1109/ICDCS.1996.507968","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.507968","url":null,"abstract":"We present a group routing protocol for a network of processes. The task of the protocol is to route data messages to each member of a process group. To this end, a tree of processes is constructed in the network, ensuring each group member is included in the tree. To build this tree, the group routing protocol relies upon the unicast routing tables of each process. Thus, group routing is a composition of a unicast routing protocol, whose detailed behavior is unknown but its basic properties are given, and a protocol that builds a group tree based upon the unicast routing tables. The design of the group routing protocol is presented in three steps. First, a basic group routing protocol is presented and proven correct. Then, the protocol is refined twice, strengthening its properties with each refinement. The final protocol has the property of adapting the group tree to changes in the unicast routing tables without compromising the integrity of the group tree, even in the presence of unicast routing loops.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131619165","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.507925
Steven P. Ketchpel, H. Garcia-Molina
In a distributed environment where nodes are independently motivated, many transactions or commercial exchanges may be stymied due to a lack of trust between the participants. The addition of trusted intermediaries may facilitate some exchanges, but others are still problematic. We introduce a language for specifying these commercial exchange problems, and sequencing graphs, a formalism for determining whether a given exchange may occur We also present an algorithm for generating a feasible execution sequence of pairwise exchanges between parties (when it exists). Indemnities may be offered to facilitate previously infeasible transactions. We show when and how they enable commercial transactions.
{"title":"Making trust explicit in distributed commerce transactions","authors":"Steven P. Ketchpel, H. Garcia-Molina","doi":"10.1109/ICDCS.1996.507925","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.507925","url":null,"abstract":"In a distributed environment where nodes are independently motivated, many transactions or commercial exchanges may be stymied due to a lack of trust between the participants. The addition of trusted intermediaries may facilitate some exchanges, but others are still problematic. We introduce a language for specifying these commercial exchange problems, and sequencing graphs, a formalism for determining whether a given exchange may occur We also present an algorithm for generating a feasible execution sequence of pairwise exchanges between parties (when it exists). Indemnities may be offered to facilitate previously infeasible transactions. We show when and how they enable commercial transactions.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129930507","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.508001
K. Tan, J. Yu
In a wireless environment, information is broadcast on communication channels to clients using powerful, battery-operated palmtops. To conserve the usage of energy, the information to be broadcast must be organized so that the client can selectively tune in at the desirable portion of the broadcast. Most of the existing work focus on uniform broadcast. However very often, a small amount of information is more frequently accessed by a large number of clients while the remainder are less in demand. This nonuniform access pattern poses several new issues. In this paper we examine these issues and look at how a nonuniform broadcast can be organized for selective tuning by the clients. We propose several new indexing schemes to facilitate selective tuning. A performance study is conducted to study and demonstrate the effectiveness of the proposed schemes.
{"title":"Energy efficient filtering of nonuniform broadcast","authors":"K. Tan, J. Yu","doi":"10.1109/ICDCS.1996.508001","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.508001","url":null,"abstract":"In a wireless environment, information is broadcast on communication channels to clients using powerful, battery-operated palmtops. To conserve the usage of energy, the information to be broadcast must be organized so that the client can selectively tune in at the desirable portion of the broadcast. Most of the existing work focus on uniform broadcast. However very often, a small amount of information is more frequently accessed by a large number of clients while the remainder are less in demand. This nonuniform access pattern poses several new issues. In this paper we examine these issues and look at how a nonuniform broadcast can be organized for selective tuning by the clients. We propose several new indexing schemes to facilitate selective tuning. A performance study is conducted to study and demonstrate the effectiveness of the proposed schemes.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131023862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.508025
T. Nakajima, Akihiro Hokimoto, K. Kurihara
In this paper, we propose software architecture using a notion of service proxies, that is a new framework for constructing applications in mobile computing environment. Our framework especially takes into account applications accessing services on Internet such as WWW browsers and MBONE conference tools. In our framework, an application is partitioned into two pieces, one piece runs on a mobile computer, and another piece runs on a stationary computer, where they are connected by wireless networks that may be replaced on the fly. The piece on a stationary computer is called a service proxy. The service proxy filters or caches data from servers before transmitting the data to the piece on a mobile computer. These two pieces are constructed by composing small objects whose composition can be dynamically reconfigured by adding or removing replaceable devices. The paper also presents a toolkit using our framework for building mobile applications easily.
{"title":"An approach for constructing mobile applications using service proxies","authors":"T. Nakajima, Akihiro Hokimoto, K. Kurihara","doi":"10.1109/ICDCS.1996.508025","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.508025","url":null,"abstract":"In this paper, we propose software architecture using a notion of service proxies, that is a new framework for constructing applications in mobile computing environment. Our framework especially takes into account applications accessing services on Internet such as WWW browsers and MBONE conference tools. In our framework, an application is partitioned into two pieces, one piece runs on a mobile computer, and another piece runs on a stationary computer, where they are connected by wireless networks that may be replaced on the fly. The piece on a stationary computer is called a service proxy. The service proxy filters or caches data from servers before transmitting the data to the piece on a mobile computer. These two pieces are constructed by composing small objects whose composition can be dynamically reconfigured by adding or removing replaceable devices. The paper also presents a toolkit using our framework for building mobile applications easily.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133386872","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1996-05-27DOI: 10.1109/ICDCS.1996.508024
S. Banerjee, Panos K. Chrysanthis
Eliminating interference between concurrently executing activities through mutual exclusion is one of the most fundamental problems in computer systems. The problem of mutual exclusion in a distributed system is especially interesting owing to the lack of global knowledge in the presence of variable communication delays. In this paper, a new token-based distributed mutual exclusion algorithm is proposed. The algorithm incurs approximately three messages at high loads, irrespective of the number of nodes N in the system. At low loads, it requires approximately N messages. The paper also addresses failure recovery issues, such as token loss.
{"title":"A new token passing distributed mutual exclusion algorithm","authors":"S. Banerjee, Panos K. Chrysanthis","doi":"10.1109/ICDCS.1996.508024","DOIUrl":"https://doi.org/10.1109/ICDCS.1996.508024","url":null,"abstract":"Eliminating interference between concurrently executing activities through mutual exclusion is one of the most fundamental problems in computer systems. The problem of mutual exclusion in a distributed system is especially interesting owing to the lack of global knowledge in the presence of variable communication delays. In this paper, a new token-based distributed mutual exclusion algorithm is proposed. The algorithm incurs approximately three messages at high loads, irrespective of the number of nodes N in the system. At low loads, it requires approximately N messages. The paper also addresses failure recovery issues, such as token loss.","PeriodicalId":159322,"journal":{"name":"Proceedings of 16th International Conference on Distributed Computing Systems","volume":"153 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128274087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：