David Reed and Liba Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355--373. North-Holland Publishing Company, Amsterdam, 1981.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Alternately, many non Byzantine tolerant systems (e.g. Harp [34] and Petal [33] serialize their actions through a primary storage node, which becomes responsible for completing the update. An alternate approach to handling both problems is to have the data stored on storage nodes be immutable [50, 51]. By definition, this eliminates the difficulties of updates for existing data. In doing so, it shifts the problem up one level; an update now consists of creating a new data item and modifying the relevant name to refer to it. Decoupling the data item creation from its visibility simplifies both, ....

....now consists of creating a new data item and modifying the relevant name to refer to it. Decoupling the data item creation from its visibility simplifies both, but making the metadata service fault tolerant often brings back the same issues. Examples of systems that use this model include SWALLOW [50], Amoeba [42] and most of the recent peer to peer file systems (e.g. Past [52] CFS [12] Farsite [1] and the archival portion of Oceanstore [30] Ivy [43] provides decentralized read write access to immutable stored data in a fashion similar to our approach. Per client update logs (which are ....

D. P. Reed and L. Svobodova. SWALLOW: a distributed data storage system for a local network. International Workshop on Local Networks (Zurich, Switzerland), August 1980.

....is to suppress it, either via locking or optimistic concurrency control [17] Alternately, many systems (e.g. Harp [20] and Petal [19] serialize their actions through a primary storage node. An alternate approach to handling both problems is to have the data stored on storage nodes be immutable [29, 30]. By definition, this eliminates the difficulties of updates for existing data. In doing so, it shifts the problem up one level an update now consists of creating a new data item and modifying the relevant name to refer to it. Decoupling the data item creation from its visibility simplifies ....

....now consists of creating a new data item and modifying the relevant name to refer to it. Decoupling the data item creation from its visibility simplifies both, but making the metadata service fault tolerant often brings back the same issues. Examples of systems that use this model include SWALLOW [29], Amoeba [25] and most of the recent peer to peer filesystems (e.g. Past [31] CFS [9] Farsite [5] and the archival portion of Oceanstore [16] Our target consistency semantics (linearizability with read aborts) have been studied previously. Notably, Pierce [28] presents a protocol ....

D. P. Reed and L. Svobodova. SWALLOW: a distributed data storage system for a local network. International Workshop on Local Networks (Zurich, Switzerland), August 1980.

....from SUNDR s protocol, which is the first to implement anything resembling traditional file system semantics without trusting the storage infrastructure. A number of file systems in the past have used cryptographic storage to keep data secret in the event of a server compromise. The swallow [15] distributed file system used client side cryptography to enforce access control. Clients encrypted files before writing them to the server. Any client could read any file, but could only decrypt the file given the appropriate key. Unfortunately, one could not grant read only access to a file. An ....

David Reed and Liba Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355--373. North-Holland Publ., Amsterdam, 1981.

....from SUNDR s protocol, which is the first to implement anything resembling traditional file system semantics without trusting the storage infrastructure. A number of file systems in the past have used cryptographic storage to keep data secret in the event of a server compromise. The swallow [14] distributed file system used client side cryptography to enforce access control. Clients encrypted files before writing them to the server. Any client could read any file, but could only decrypt the file given the appropriate key. Unfortunately, one could not grant readonly access to a file. An ....

David Reed and Liba Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355--373. North-Holland Publ., Amsterdam, 1981.

....and will allow client caches to be used in the reconstruction of server state in the event of a failure. 2 Related work While a number of file systems have used cryptographic storage in the past, none has provided strong integrity guarantees in the face of a compromised server. The swallow [13] distributed file system used client side cryptography to enforce access control. Clients encrypted files before writing them to the server. Any client could read any file, but could only decrypt the file given the appropriate key. Unfortunately, one could not grant read only access to a file. An ....

D. Reed and L. Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355--373. North-Holland Publ., Amsterdam, 1981.

....Institute of Technology. The main design components, repositories and brokers are prototypically implemented and tested under artificial workload. The current status of research is unknown. Contact: Liba Svobodova, IBM Zurich Research Lab. 8803 Ruschlikon, Switzerland. References: 315] 77] [316], 317] 82] 2.50 V Main Goal The V distributed system was developed at Stanford University with the purpose of connecting multiple workstations, access and location transparency provided. A small kernel is supposed to provide high performance communication and be used by servers which offer ....

D.P. Reed and L. Svobodova, "SWALLOW: A Distributed Data Storage System for a Local Network", In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355--373, North-Holland Publishing Company, 1981. 80

....algorithms for atomic commitment, such as the two phase commmit (2PC) and three phase commit (3PC) protocols [5] for distributed databases. These sophisticated techniques are unnecessary for concurrent databases under the all or nothing assumption. The SWALLOW distributed storage system [54] stores objects in a collection of autonomous repositories. Each object is represented as a sequence of versions, called a version history, which records successive updates to the object over time. SWALLOW creates a commit record at each repository at which a transaction updates objects and links ....

David Reed and Liba Svobodova. SWALLOW: A Distributed Data Storage System for a Local Network. In Local Networks for Computer Communications: Proceedings of the IFIP Working Group 6.4 International Workshop on Local Networks, pages 355--373, Zurich, Switzerland, August 1980. North-Holland.

....Existing local files may be modified, but this feature is used only for special purposes such as updating local log files. Tools such as the editor and compiler treat local files as immutable too, by always creating new file versions when writing results to disk. The Swallow file system design [13] first explored the benefits of immutable versions. A local working directory provides the naming environment in which a programmer works. CFS prepends the current local working directory name to any file name argument that does not start with the character 4 . There is no search rule ....

Reed, D.P. and Svobodova, L., "SWALLOW: a distributed data storage system for a local network," Local Networks for Computer Communications, North-Holland, Amsterdam, 1981, pp. 355--373.

....this thesis discusses the design of SFS and the SFS protocol, a prototype imlementation of SFS, and performance measurements of the prototype implementation. 2 Related Work Many projects have explored distributed file systems in the context of caching, performance, security, or a global namespace [3, 7, 8, 11, 12, 13, 18, 21, 22, 23, 34, 39, 44, 45, 46, 48, 49, 53, 57]. While SFS s design goals have been met individually by previous projects, SFS is the first file system to provide security, a global namespace, and completely decentralized control at the same time. One of the file systems in most widespread use today is Sun s NFS [44, 54] NFS is purely a ....

David Reed and Liba Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications. North-Holland Publ., Amsterdam, 1981.

....the scalability of the system. Unfortunately, we currently do not have a JetFile port to a more popular kind of machine. 9 Related Work The idea of building a storage system around the distribution of immutable object versions is not entirely new. These ideas can be traced back to the SWALLOW [28] distributed data storage system. SWALLOW is, however, mostly concerned with the mechanisms necessary to support commitment synchronization between objects, a relatively heavyweight mechanism that is usually not required in distributed file systems. Another system based on optimistic versioned ....

D. P. Reed, L. Svobodova, SWALLOW: A Distributed Data Storage System for a Local Network, Local Networks for Computer Communications, North-Holland, Amsterdam 1981.

....a file system as an append only log is not a new idea. Several other file The LFS Storage Manager May 1, 1991 systems, motivated by the advent of write once media such as optical disks, have used similar mechanisms. Write once storage managers with random access to files include SWALLOW[8], the Optical File Cabinet[9] and others[10] These file systems were intended principally for archival storage and not as high performance file servers. Another way of viewing the LFS design is to see its roots in file systems like Cedar[11] that use logging to improve write performance and ....

D. Reed and Liba Svobodova, "SWALLOW: A Distributed Data Storage System for a Local Network," Local Networks for Computer Communications, pp. 355-373 North-Holland, (1981).

....increase the checkpoint interval. 6. Related work The log structured file system concept and the Sprite LFS design borrow ideas from many different storage management systems. File systems with log like structures have appeared in several proposals for building file systems on write once media[17, 18]. Besides writing all changes in an append only fashion, these systems maintain indexing information much like the Sprite LFS inode map and inodes for quickly locating and reading files. They differ from Sprite LFS in that the write once nature of the media made it unnecessary for the file systems ....

D. Reed and Liba Svobodova, "SWALLOW: A Distributed Data Storage System for a Local Network," Local Networks for Computer Communications, pp. 355-373 North-Holland, (1981).

....in some, but not all applications. An endto end acknowledgement is still required for applications in which the action requested of the target host should be done only if similar actions requested of other hosts are successful. This kind of application requires a two phase commit protocol[5,10,15], which is a sophisticated endto end acknowledgement. Also, if the target application may either fail or refuse to do the requested action, and thus a negative acknowledgement is a possible outcome, an end to end acknowledgement may still be a requirement. Secure transmission of data Another ....

....mechanism at a higher level than the communication subsystem must control the ordering of actions. SALTZER et al. End to End Arguments in System Design 7 Transaction management We have now applied the end to end argument in the construction of the SWALLOW distributed data storage system[15], where it leads to significant reduction in overhead. SWALLOW provides data storage servers called repositories that can be used remotely to store and retrieve data. Accessing data at a repository is done by sending it a message specifying the object to be accessed, the version, and type of ....

Reed, D.P., and Svobodova, L.: SWALLOW: A distributed data storage system for a local network. In West, A., and Janson, P., ed. Local Networks for Computer Communications, Proc. IFIP Working Group 6.4 International Workshop on Local Networks. North-Holland, Amsterdam, 1981, pp.355-373.

No context found.

David Reed and Liba Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355--373. North-Holland Publishing Company, Amsterdam, 1981.

No context found.

David P. Reed and Liba Svobodova. SWALLOW: a distributed data storage system for a local network. International Workshop on Local Networks (Zurich, Switzerland), August 1980.

No context found.

David P. Reed and Liba Svobodova. SWALLOW: a distributed data storage system for a local network. International Workshop on Local Networks (Zurich, Switzerland), August 1980.

No context found.

David Reed and Liba Svobodova. Swallow: A distributed data storage system for a local network. In A. West and P. Janson, editors, Local Networks for Computer Communications, pages 355-- 373. North-Holland Publ., Amsterdam, 1981.

No context found.

D. P. Reed and L. Svobodova, "SWALLOW: A Distributed Data Storage System for a Local Network," Int'l. Wksp. Local Networks, Zurich, Switzerland, Aug. 1980.

No context found.

Reed, D. and Svobodova, L. SWALLOW: A Distributed Data Storage System for a Local Network. Local Networks for Computer Communications, North Holland (1981) 355-373.

No context found.

Reed, D.P., and Svobodova, L.: SWALLOW: A distributed data storage system for a local network. In West, A., and Janson, P., ed. Local Networks for Computer Communications, Proc. IFIP Working Group 6.4 International Workshop on Local Networks. North-Holland, Amsterdam, 1981, pp.355-373.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC