V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. Proc. of 25th VLDB Conference, pages 709--720, 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....to reoptimize parts of queries after blocking operators [24] There is also a lot of work on adaptive query operators, an area we believe to be relevant to sensor networks. Examples include work on memory adaptive sorting and hashing [13, 28, 30, 34, 53, 54] and online aggregation algorithms [15, 18, 39, 48]. Eddies push the idea of feedback on a tuple by tuple basis in online aggregation to adapting join orders at the same frequency [4] Other relevant work includes sequence query processing [42, 43] and temporal and time series databases [52] 4.5 Related Projects We conclude our discussion of ....

V. Raman, B. Raman, and J. M. Hellerstein. Online dynamic reordering for interactive data processing. In M.P. Atkinson, M. E. Orlowska, P. Valduriez, S. B. Zdonik, and M. L. Brodie, editors, VLDB'99, Proceedings of 25th International Conference on Very Large Data Bases, September 7-10, 1999.

....data. The idea of improving the efciency of merges over very large databases using windows was studied by DeWitt et al. 2] A key component in merging distributed data streams is sorting the two streams by their common keys. Previous work on sorting data streams was conducted by Juggle [19]. In this paper they outline a method by which data can be retrieved from a distributed data repository and served in a partially ordered fashion. Assumptions of the data being indexed are made. Their focus is the interactive control of queries to streaming data. Our focus is on non interactive ....

V. Raman, B. Raman, and J. Hellerstein, Online Dynamic Reordering for Interactive Data Processing In Proc. of the 1999.

....found that VC improved the delivery of the important results in types of query plans for which it was designed. When used in conjunction with SJP, VC increased the number of important tuples produced by SJP about 50 during the early stages of query execution. 5 Related Work The Juggle operator [RRH99] is a pipelined best e ort reordering operator whose goal is also to produce important results faster. It takes an unordered set of tuples and produces a result that is nearly sorted. Input tuples are bu ered in a sorted list which is updated as new tuples arrive. When the parent operator ....

V. Raman, B. Raman, J. M. Hellerstein. Online Dynamic Reordering for Interactive Data Processing. VLDB Conf., 1999.

....incorporated into our framework, because it would be unpractical to have all tuples involved in a query transit, potentially many times, through a single LeSelect server. Our work is also related to existing results on improving the output rate for part of the query result, as described, e.g. in [18, 20]. However, our work is di erent in that we used parallelism, asynchronism and duplicates to improve the output rate of the distributed BindJoin operator. As directions of future work, we plan to extend our framework to sharing blob and function cache among several queries, and to processing of ....

A. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. In Proc. of the VLDB Conf., 1999.

.... are available at the software web page [39] 2 XYZ Architecture The main components of the XYZ architecture (Figure 2) are a Data Source,aTransformation Engine that applies transforms along 2 paths, an Online Reorderer to support interactive scrolling and sorting at the user interface [42, 41], and an Automatic Discrepancy Detector. We proceed to discuss these in turn. 2.1 Data Source XYZ accepts input data as a single, pre merged input stream. This input can come from an ODBC source or any ASCII file descriptor. files, pipes, etc. The ODBC source can be used to access data from ....

.... Discrepancy Detection Spreadsheet Display specify undo transforms Online reorderer scroll check for anomalies get page scrollbar position Side Disk Figure 2: XYZ Architecture (such as sensor feeds) The interface supports this behavior using the Online Reorderer developed in [42]. This reordered continually fetches tuples from the source and divides them into buckets based on a (dynamically computed) histogram on the sort column, spooling them to a side disk if needed. When the user scrolls to a new region, the reorderer picks a sample of tuples from the bucket ....

V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. In Proc. Intl. Conference on Very Large Data Bases, 1999.

No context found.

V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. In Proc. Intl. Conference on Very Large Data Bases, 1999.

....a particular machine (see Figure 2) and a insert(tuple T, int PartID) method for queuing up a tuple for the destination PartID. The buffer is implemented as a hashtable keyed on the destination of the tuple, and pins down a fixed amount of memory for that hashtable. It is similar to the Juggle [22] operator, except it does not spill tuples to disk. Flux Prod drains tuples from the buffer for only the connections that can accept another tuple. Flux Prod first calls getNext( on the producer instance if the buffer has space to hold another tuple, inserts the returned tuple into the buffer ....

V. Raman, B. Raman, and J. M. Hellerstein. Online Dynamic Reordering for Interactive Data Processing. In VLDB, 1999.

....the plan on a continuous basis while a query is running. Eddies are modules that adaptively decide how to route data to other query operators on a tuple by tuple basis [AH00] choosing orderings among commutative modules. Juggle performs online reordering for prioritizing records by content [RRH99]. Flux routes tuples among machines in a cluster to support parallelism with load balancing and fault tolerance [SCHF03] Architecturally, these modules are indistinguishable from the other more traditional modules: they simply consume and produce records via the usual Fjords API. However, these ....

....routing policies on disk access behavior. Queries accessing data that spans memory and disk also raise significant Quality of Service issues, in terms of deciding what work to drop when the system is in danger of falling behind the incoming data stream. Our earlier work on the Juggle operator [RRH99] and on dynamic pipeline processing [UF02] provide mechanisms for pushing user preferences down into the query execution process. Such techniques will need to be integrated into TelegraphCQ. Egress Modules. Analogous to our ingress modules, we also plan to investigate mechanisms for managing and ....

Raman, V., Raman, B., and Hellerstein, J., Online Dynamic Reordering for Interactive Data Processing. In VLDB (1999).

....a particular machine (see Figure 2) and a insert(tuple T, int PartID) method for queuing up a tuple for the destination PartID. The buffer is implemented as a hashtable keyed on the destination of the tuple, and pins down a fixed amount of memory for that hashtable. It is similar to the Juggle [22] operator, except it does not spill tuples to disk. Flux Prod drains tuples from the buffer for only the connections that can accept another tuple. Flux Prod first calls getNext( on the producer instance if the buffer has space to hold another tuple, inserts the returned tuple into the buffer ....

V. Raman, B. Raman, and J. M. Hellerstein. Online Dynamic Reordering for Interactive Data Processing. In VLDB, 1999.

....way that balances this tradeoff. 2 Potter s Wheel Architecture The main components of the Potter s Wheel architecture (Figure 2) are a Data Source,aTransformation Engine that applies transforms along 2 paths, an Online Reorderer to support interactive scrolling and sorting at the user interface [23, 21], and an Automatic Discrepancy Detector. 2.1 Data Source Potter s Wheel accepts input data as a single, pre merged stream, that can come from an ODBC source or any ASCII file descriptor (or pipe) The ODBC source can be used to query data from DBMSs, or even from distributed sources via ....

....the user starts Potter s Wheel on a dataset, the spreadsheet interface appears immediately, without waiting until the input has been completely read. This is important when transforming large datasets or never ending data streams. The interface supports this behavior using an Online Reorderer [23] that continually fetches tuples from the source and divides them into buckets based on a (dynamically computed) histogram on the sort column, spooling them to disk if needed. When the user scrolls to a new region, the reorderer picks a sample of tuples from the bucket corresponding to the ....

V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. In VLDB, 1999.

....Section 6 and conclude in Section 7. 2 Potter s Wheel Architecture The main components of Potter s Wheel architecture (Figure 2) are a Data Source,aTransformation Engine that applies transforms along 2 paths, an Online Reorderer to support interactive scrolling and sorting at the user interface [35, 34], and an Incremental Discrepancy Detector. We proceed to discuss these in turn. 2.1 Transformation Engine Transforms specified by the user need to be applied in two places. First, they need to be applied to records visible on screen. With the spreadsheet user interface this is done when the user ....

....screen. 2.3 Interface used for Displaying Data Our user interface is a Scalable Spreadsheet [34] that allows users to interactively re sort on any column, and scroll in a representative sample of the data, even over large datasets. The interface supports this behavior using an Online Reorderer [35] that continually fetches tuples from the source and divides them into buckets based on a (dynamically computed) histogram on the sort column, spooling them to a side disk if needed. When the user scrolls to a new region, the reorderer picks a sample of tuples from the bucket corresponding to the ....

V. Raman, B. Raman, and J. M. Hellerstein. Online dynamic reordering for interactive data processing. In VLDB, 1999.

....disks, and the rate of production from various partitions may change over time depending on performance characteristics and utilization of the different disks. Finally, Online Aggregation systems explicitly allow users to control the order in which tuples are delivered based on data preferences [RRH99] resulting in similar effects. 1.2 Architectural Assumptions Telegraph is intended to efficiently and flexibly provide both distributed query processing across sites in the wide area, and parallel query processing in a large shared nothing cluster. In this paper we narrow our focus somewhat to ....

V. Raman, B. Raman, and J. M. Hellerstein. Online Dynamic Reordering for Interactive Data Processing. In Proc. 25th International Conference on Very Large Data Bases (VLDB), pages 709--720, Edinburgh, 1999.

....ffl User Interface Complexity: In large scale systems, many queries can run for a very long time. As a result, there is interest in Online Aggregation and other techniques that allow users to Control properties of queries while they execute, based on refining approximate results [HAC 99, RRH99] For all of these reasons, we expect query processing parameters to change significantly over time in Telegraph, typically many times during the execution of a single query. As a result, it is not appropriate to use the traditional architecture of optimizing a query and then executing a static ....

....partitions and hence from different value ranges may change over time depending on performance characteristics and utilization of the different disks. Finally, Online Aggregation systems explicitly allow users to control the order in which tuples are delivered based on data preferences [RRH99] this can have a similar effect on query predicates that are correlated with the user controlled preferences. 1.2 Architectural Assumptions Telegraph is intended to efficiently and flexibly provide both distributed query processing across sites in the wide area, and parallel query processing ....

Vijayshankar Raman, Bhaskaran Raman, and Joseph M. Hellerstein. Online Dynamic Reordering for Interactive Data Processing. In Proc. 25th International Conference on Very Large Data Bases (VLDB), pages 709--720, Edinburgh, 1999.

....position. While she is exploring the data fetched so far, we fetch more items (from the source) and sort them in the background. We use a (dynamically computed) equi depth histogram over the data to decide which items need to be returned at a given scrollbar position, and Online Reordering [RRH99] for sorting the tuples while the user is looking at the rest of the data. Briefly, we prefetch items from the source and build an approximate hash index on disk based on the histogram buckets while the user is scanning the data that has been fetched so far. To avoid confusing the user, we do not ....

V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. In VLDB, 1999. (to appear).

No context found.

V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. Proc. of 25th VLDB Conference, pages 709--720, 1999.

No context found.

V. Raman, B. Raman, and J. M. Hellerstein. Online Dynamic Reordering for Interactive Data Processing. In VLDB 1999.

No context found.

V. Raman, et al.. Online dynamic reordering for interactive data processing. In The VLDB Journal , pp. 709--720. 1999.

No context found.

V. Raman, et al.. Online dynamic reordering for interactive data processing. In The VLDB Journal, pp. 709--720. 1999.

No context found.

V. Raman, B. Raman, and J. Hellerstein, Online Dynamic Reordering for Interactive Data Processing In Proc. of the 1999.

No context found.

V. Raman, B. Raman, and J. Hellerstein, Online Dynamic Reordering for Interactive Data Processing In Proc. of the 1999.

No context found.

V. Raman, et al.. Online dynamic reordering for interactive data processing. In The VLDB Journal , pp. 709--720. 1999.

No context found.

V. Raman, B. Raman, and J. Hellerstein. Online dynamic reordering for interactive data processing. Proc. of 25th VLDB Conference, pages 709--720, 1999.

No context found.

V. Raman, B. Raman, and J. Hellerstein, Online Dynamic Reordering for Interactive Data Processing In Proc. of the 1999.

No context found.

V. Raman, B. Raman, J. M. Hellerstein, "Online Dynamic Reordering for Interactive Data Processing", Proceedings of the 1999 VLDB Conference, Edinburgh, Scotland, September 1999.

No context found.

V. Raman, B. Raman, and J.M. Hellerstein, "Online Dynamic Reordering for Interactive Data Processing," to be published in Proc. 25th Int'l Conf. Very Large Data Bases, Morgan Kaufmann, San Francisco, 1999.

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