D. G. Severance and G. M. Lohman, "Differential Files: Their Applications to the Maintenance of Large Databases," ACM TODS 1 (3), 1976, 256--367.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....to the disk many more times than the no force policy. In general, the steal no force combination seems the appropriate choice for practical systems [43] The other mechanisms proposed in the literature for recovery in centralized systems include shadow paging [38] and the differential files [44]. However, due to some inherent problems in these techniques, commercial systems usually avoid them and instead prefer to implement the logging mechanisms [43] A study of the performance issues that arise in the integration of various concurrency control protocols with the recovery protocols is ....

D. G. Severance and G. M. Lohman, "Differential Files: Their Application to the Maintenance of Large Databases", Transactions On Database Systems

.... before another transaction can dirty it [Gra96a] In general, the steal no force combination seems the appropriate choice for practical systems [GR93] The other mechanisms proposed in the literature for recovery in centralized systems include shadow paging [LS76] and the differential files [SL76] However, due to some inherent problems in these techniques, commercial systems usually avoid them and instead prefer to implement the logging mechanisms [GR93] A study of the performance issues that arise in the integration of various concurrency control protocols with the recovery protocols ....

D. G. Severance and G. M. Lohman. Differential Files: Their Application to the Maintenance of Large Databases. ACM Transactions on Database Systems, 1(3), 1976.

....of the DSM copy. Hence, given high update rates, the overhead of maintaining the mirrors up to date may result in a serious performance penalty. The solution we are implementing uses an intermediate representation of the relation to serve as a differential file to record updates and inserts [17]. The differential file is implemented as a relation with three attributes having the schema (Tuple Id, Attribute Id, Value) A single entry represents a new attribute value of the original tuple. An insert operation would now result in the insert of n tuples to this relation, the main difference ....

D.Severance, G.Lohman. Differential Files: Their application to the maintenance of large databases. ACM TODS September 1976 vol. 1 number 3.

....an atomic manner. The three popular concurrency control methods are locking [1, 2] timestamping [3] and optimistic algorithms [4] which have been proposed for both centralized and distributed environments. In order to commit a transaction, logging [2] shadow paging [5] and differential files [6] techniques are proposed for centralized database systems, whereas two phase commit (2PC) protocol [2] and three phase commit (3PC) protocol [7] are proposed for DDBSs. The general aim of a transaction processing (concurrency control and commit) approach is to ensure the consistency of the ....

D.G.Severance, and G.M.Lohman, Differential files: Their application to the maintenance of large databases, ACM Transactions on Database Systems, 1(3), September 1976, 256-267.

....changes between versions of data was formalized as stringto string correction with block move [24] detecting maximally matching regions of a file at an arbitrarily fine granularity without alignment. However, delta compression continued to rely on the alignment of data, as in database records [23], and the grouping of data into block or line granularity, as in source code control systems [22, 25] to simplify the combinatorial task of finding the common and different strings between versions. Efforts to generalize delta compression to un aligned data and to minimize the granularity of the ....

D. G. Severance and G. M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Transactions on Database Systems, 1(2):256--267, September 1976.

....projection views. Hanson [Han87b, Han87a] reviews the differential view update methods developed by Blakeley (described above) the hypothetical relations implementation work of Woodfill Stonebraker [WS83] and of Agrawal Dewitt [AD83] and the differential file techniques of Severance [SL76] Hanson also discusses the implementation of views containing aggregate queries (SUM, COUNT, etc. As has been remarked, this work employs similar differential view update techniques for the update of materialized sample views. Hanson [Han87a] also considers a RETE network approach to ....

D.G. Severance and G.M. Lohman. Differential Files: Their Application to the Maintenance of Large Databases. ACM Transactions on Database Systems, 1(3):256--267, September 1976.

....of size F . These bit patterns are OR ed to form the signature. Word Signature data 001 000 110 010 base 000 010 101 001 doc. signature 001 010 111 011 Ideally, half of the bits should be 1 in a doc. signature. Sti60] Used for spelling checking in UNIX [McI82] in differential files [SL76] in text retrieval in CM 2 [Sta88] etc. 36 5.4 Clustering Dominating method in information retrieval. Sal71] SM83] VR79] Well suited for keyword queries . 0 0 0.6 zoo data Aaron indexing . data . document 37 Main idea: group similar vectors together; group ....

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM TODS, 1(3):256--267, September 1976.

....do not support persistent deltas, nor constructs to create, manipulate and compare completely unrelated deltas. Furthermore, as detailed in Subsection 6.4 below, Heraclitus can be used directly to specify, and thereby implement, novel execution models for specialized purposes. Differential files [SL76] are a low level implementation technique that support efficient representation of multiple versions of a database. Unlike differential files, deltas in the Heraclitus framework are manipulated directly by constructs supported by the user level programming language. Furthermore, we support a ....

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Trans. on Database Systems, 1(3), September 1976.

....(circled with dotted lines) We detect the common subtrees at compile time and replace them with the final result of the leftmost subtree in order to avoid duplicate computations when executing the tree 7 . 6 Related Work There are two fundamental differences between differential files [SL76] and Heraclitus[Alg] First, the main focus of Heraclitus is at the conceptual and programming level as opposed to the phys 7 We are assuming that a query tree is executed in a post traversal order. 13 (b) a) D1 E D3 D2 when when when D3 D2 when when D3 when D1 E Figure 2: ....

D. Severance and G. Lohman. Differential files: Their application to the maintenance of large databases. ACM Trans. on Database Systems, 1(3):256--267, September 1976.

....day, year) The symbol R denotes R as of current time. We assume that all tuples are augmented with immutable and globally unique tuple identifiers. Consequently, each tuple is uniquely identified even though duplicates may occur. Changes made to base relations are stored on differential files [SL76] from which insertion sets and deletion sets are extracted. The symbol R t 1 Gammat 2 I denotes the insertion set of R relative to the time interval (t 1 ; t 2 ] The symbol R t 1 Gammat 2 D denotes the deletion set of R relative to the time interval (t 1 ; t 2 ] We assume that insertion ....

D.G. Severance and G.M. Lohman. Differential Files: Their Application to the Maintenance of Large Databases. ACM Transactions on Database Systems, 1(3):256-- 267, September 1976.

.... databases make use of many reliability features such as stable storage, transactions, nested transactions [76] commit and recovery protocols [103] nonblocking commit protocols [102] termination protocols [104] checkpointing, replication, primary backups, logs audit trails, differential files [99], and timeouts to detect failures. Operating system support is required to make these mechanisms more efficient [47, 87, 119, 131] One aspect of reliability not stressed enough in DCS research is the need for robust solutions, i.e. the solutions must explicitly assume an unreliable network, ....

....because it integrates concurrency control with recovery. In this scheme, each change to the database results in a new version. Concurrency control is enforced on a version basis and old versions serve as checkpoints. Handling multiple versions efficiently is accomplished by differential files [99]. Validation is a technique which permits unrestricted access to data items (resulting in no blocking and hence fast access to data) but then checks for potential conflicts at the commit point. The commit point is the time at which a transaction is sure that it will complete. This approach is ....

D.G. Severance and G.M. Lohman, "Differential Files: Their Application to the Maintenance of Large Databases," ACM Transactions on Database Systems, Vol. 1, No. 3, September 1976.

....problems for our application. The major idea behind both solutions is to use a signature approach, that is, to create a filter that will allow some false hits, but no false dismissals. This philosophy has been successfully used in several environments (text retrieval [14] differential files [42], hash join algorithms, spelling checking [29] etc. Thus, in both the above problems, our goal is to find a mapping of the feature vector X into a vector X 0 = f( X) where X 0 is a vector in a more suitable space, with a distance function D 0 ( which will underestimate ....

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM TODS, 1(3):256--267, September 1976.

.... Rosenkrantz and Yu demonstrate a reconstruct algorithm requiring O(L) steps by computing a special form of the insert delete delta [22] Severance and Lohman show how to use file deltas for the storage of versions of large objects in databases and, more generally, for implementing transactions [15]. Alderson presents a space efficient technique based on B trees and file deltas that allows partial records to be retrieved without fully constructing the entire record, and for which the get operation has a predictable time complexity for all versions [1] Fraser and Myers demonstrate an editor ....

Severance, D. S., and Lohman, G. M. Differential files: Their application to the maintenance of large databases. ACM Transactions on Database Systems (Sept. 1976), 256--267.

....(increments) of the modifications are propagated to the client s site. The set of algorithms that carry out these tasks are based on the Incremental Access Methods for relational operators described in [15] and involve looking at the update logs of the server and transmitting differential files [21]. This significantly reduces data transmission over the network as it only transmits the increments affecting the bound object, compared with traditional CS architectures in which query results are continuously transmitted in their entirety. It is important to point out some of the ....

D. G. Severance and G. M. Lohman. Differential Files: Their Application to the Maintenance of Large Databases. ACM--Transactions on Database Systems, 1(3), September 1976.

....is to extend active database techniques, so that rule conditions can explicitly access multiple deltas corresponding to the different proposed updates. 2.3 Related techniques This section concludes with a brief comparison of the Heraclitus paradigm with related techniques. Differential files [SL76] are a low level implementation technique that support efficient representation of multiple versions of a database. Unlike differential files, deltas in the Heraclitus framework are manipulated directly by constructs in the user level programming language. Furthermore, we support a family of ....

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Trans. on Database Systems, 1(3), September 1976.

....complicated in the case of Stepped Hash. For lack of space we do not describe either: see the full version of the paper for details. 6 Discussion Bloom filters (bitmap filters) can be used to avoid looking up many of the runs that do not contain any records for a query key, as is done in, e.g. SL76] The direct insert algorithm clearly benefits from a parallel disk system, since such a system supports a larger number of seeks per second. Parallel I O can also be used with our techniques. The output runs or buckets can be striped across multiple disks, so that they transfer data out in ....

....a small increase in look up cost. Stepped Merge has a slight edge in terms of insertion cost over Stepped Hash and a considerable benefit in terms of look up cost. 8 Related Work The idea of maintaining a log of recent changes separately from the main data file is quite old; see for example [SL76] which discusses differential files. The idea of using Bloom filters has also been explored in [SL76] However, their goal was not to save I O as compared to standard structures like B trees, but to avoid changing the main file. They do not consider issues of multi level organization of ....

[Article contains additional citation context not shown here]

D.G. Severance and G.M Lohman. Differential files: Their applications to the maintenance of large databases. ACM Transactions on Database Systems, 1(3):256--367, September 1976.

....is to extend active database techniques, so that rule conditions can explicitly access multiple deltas corresponding to the different proposed updates. 2.3 Related techniques This section concludes with a brief comparison of the Heraclitus paradigm with related techniques. Differential files [SL76] are a low level implementation technique that support efficient representation of multiple versions of a database. Unlike differential files, deltas in the Heraclitus framework are manipulated directly by constructs in the user level programming language. Furthermore, we support a family of ....

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Trans. on Database Systems, 1(3), September 1976.

....to detecting changed regions of a file at an arbitrarily fine granularity and alignment. Even though the general problem of detecting and encoding version to version modifications was well defined, delta compression applications continued to rely on the alignment of data, as in database records [13], and the grouping of data into block or line granularity, as in source code control systems [12, 15] to limit the amount of required metadata and simplify the combinatorial task of finding the common and different strings between files. Efforts to generalize delta compression to data that are ....

Dennis G. Severance and Guy M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Transactions on Database Systems, 1(2):256--267, September 1976.

....as they occur; 3) Early: Here the updates are propagated at any time before the coherency conditions are violated; 4) Delayed update: Here the updates are propagated so that no condition is violated. The values are installed whenever it is convenient. Differential file algorithm is analysed in [Sev76]. A differential file is used to record the changes made on the primary copy. This differential file is then used to update the copies. The time of update depends on the algorithm. For instance, it can be on time of access of replica, on user demand, or done periodically. Differential refresh ....

Severance, D.G., Lohman, G., Differential Files: Their Application To The Maintenance of Large Databases, ACM Trans. on Database systems, Vol 1, Iss 3, Sep. 1976.

....maximally matching regions of a file at an arbitrarily fine granularity without alignment. Even though the general problem of detecting and encoding version to version modifications was well defined, delta compression applications continued to rely on the alignment of data, as in database records [13], and the grouping of data into block or line granularity, as in source code control systems [12, 15] to simplify the combinatorial task of finding the common and different strings between files. Efforts to generalize delta compression to data that are not aligned and to minimize the granularity ....

D. G. Severance and G. M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Transactions on Database Systems, 1(2):256--267, September 1976.

....we have to store the fl kopt triplets of the form (row, column, delta) for the outlier cells. This should be done in a hash table, where the key is the combination of (row M column) that is, the order of the cell in the row major scanning. Optionally, we could use a main memory Bloom filter [22], which would predict the majority of non outliers, and thus save several probes into the hash table. Reconstruction: Reconstructing the value of a single cell, say (i; j) now requires: ffl one disk access to fetch the i th row of U (as in plain SVD) and then k opt main memory operations with ....

....benefits of each alternative. 6.2 Practical Issue In one of our real datasets (results not reported in this paper) we observed that there were several customers that did not make any purchases at all. An engineering solution in this case is to flag all these customers, and build a Bloom filter [22], to help detect them quickly. 7 Conclusions We have examined the problem of providing fast random access capability on a huge collection of time sequences. Our contributions are: ffl The formulation of the problem as a lossy compression problem, and the proposal of several tools to solve it, ....

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM TODS, 1(3):256--267, September 1976.

....schemes such as extendible hashing [9] and a myriad others. However, there are a few access methods which migrate their entries from one segment to another: MD OD RTrees of Kolovson and Stonebraker ( 15] and Time Split B trees of Lomet and Salzberg ( 17] 18] The Differential File approach [25] also collects up changes in a small component, later performing updates to the full sized structure. We will consider these structures in a bit more depth. 27 First of all we should analyze exactly why the LSM tree beats the Continuum Structure in terms of I O performance, reducing the disk ....

....correspondence between the LSM Tree and the MD OD R Tree if the latter is promoted up one level of the memory hierarchy to use memory and disk, but most of the details are not the same because of the differences in the features of the three media. Differential File The Differential File approach [25] starts with a main data file which remains unchanged over an extended period, while newly added records are placed into a specific overflow area known as a Differential File. At some future point (not carefully specified) it is assumed that the changes will be amalgamated with the main data file, ....

Dennis G. Severance and Guy M. Lohman, "Differential Files: Their Application to the Maintenance of Large Databases", ACM Trans. on Database Systems, V 1, N 3 (Sept. 1976), pp 256-267.

....granularity. For file systems, there are no guarantees that modifications have page alignment. While dirty bits are effective for databases, they may not apply well to file system backup. To improve on the granularity of backup, logging methods have been used to record the changes to a database [9, 14] and a file system [16] A logging system records every write during an epoch to a log file. This can be used to recover the version as it existed at the start of an epoch into its current state. While semantically similar to delta compression, logging does not provide the compressibility ....

Dennis G. Severance and Guy M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Transactions on Database Systems, 1(2):256--267, September 1976.

No context found.

D. G. Severance and G. M. Lohman, "Differential Files: Their Applications to the Maintenance of Large Databases," ACM TODS 1 (3), 1976, 256--367.

No context found.

D.G. Severance and G.M. Lohman. Differential files: Their application to the maintenance of large databases. ACM Trans. on Database Systems, 1(3), September 1976.

First 50 documents

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