(I) Experiment of the necessity of TimeseriesMetadata
After we store TimeseriesMetadata together with ChunkMetadata, the necessity of TimeseriesMetadata needs to be reconsidered. We need some experiments for decision.
TimeseriesMetadata for Aggregation query and raw data query under different circumstances for one timeseries in one tsfile.
Each chunk has 100 points. Each query contains 500 TsFiles.
(1) with TimeseriesMetadata: origin TimeseriesMetadata
(2) without TimeseriesMetadata: TimeseriesMetadata has no statistics
And test query for 1 timeseries in TsFile which have 1 timeseries and 1000 timeseries seperately.
Writing:
String path = "/home/fit/szs/data/data/sequence/root.sg/0/" + chunkNum + "/test" + fileIndex + ".tsfile"; File f = FSFactoryProducer.getFSFactory().getFile(path); if (f.exists()) { f.delete(); } try (TsFileWriter tsFileWriter = new TsFileWriter(f)) { // only one timeseries tsFileWriter.registerTimeseries( new Path(Constant.DEVICE_PREFIX, Constant.SENSOR_1), new UnaryMeasurementSchema(Constant.SENSOR_1, TSDataType.INT64, TSEncoding.RLE)); // construct TSRecord for (int i = 1; i <= chunkNum * 100; i++) { TSRecord tsRecord = new TSRecord(i, Constant.DEVICE_PREFIX); DataPoint dPoint1 = new LongDataPoint(Constant.SENSOR_1, i); tsRecord.addTuple(dPoint1); // write TSRecord tsFileWriter.write(tsRecord); if (i % 100 == 0) { tsFileWriter.flushAllChunkGroups(); } } }
Raw data query:
for (int fileIndex = 0; fileIndex < fileNum; fileIndex++) { // file path String path = "/home/fit/szs/data/data/sequence/root.sg/0/" + chunkNum + "/test" + fileIndex + ".tsfile"; // raw data query try (TsFileSequenceReader reader = new TsFileSequenceReader(path); ReadOnlyTsFile readTsFile = new ReadOnlyTsFile(reader)) { ArrayList<Path> paths = new ArrayList<>(); paths.add(new Path(DEVICE1, "sensor_1")); QueryExpression queryExpression = QueryExpression.create(paths, null); long startTime = System.nanoTime(); QueryDataSet queryDataSet = readTsFile.query(queryExpression); while (queryDataSet.hasNext()) { queryDataSet.next(); } costTime += (System.nanoTime() - startTime); } }
Aggregation query:
long totalStartTime = System.nanoTime(); for (int fileIndex = 0; fileIndex < fileNum; fileIndex++) { // file path String path = "/home/fit/szs/data/data/sequence/root.sg/0/" + chunkNum + "/test" + fileIndex + ".tsfile"; // aggregation query try (TsFileSequenceReader reader = new TsFileSequenceReader(path)) { Path seriesPath = new Path(DEVICE1, "sensor_1"); long startTime = System.nanoTime(); TimeseriesMetadata timeseriesMetadata = reader.readTimeseriesMetadata(seriesPath, false); long count = timeseriesMetadata.getStatistics().getCount(); costTime += (System.nanoTime() - startTime); } } System.out.println( "Total raw read cost time: " + (System.nanoTime() - totalStartTime) / 1000_000 + "ms"); System.out.println("Index area cost time: " + costTime / 1000_000 + "ms");
1 timeseries in one tsfile:
chunk number | 1 | 2 | 3 | 5 | 8 | 10 | 15 | 20 | 25 | ||
raw | with timeseriesMetadata | overall cost time (ms) | 210 | 230 | 237 | 250 | 276 | 297 | 309 | 344 | 374 |
index area time (ms) | 116 | 131 | 142 | 156 | 185 | 197 | 220 | 255 | 282 | ||
without timeseriesMetadata | overall cost time (ms) | 219 | 223 | 242 | 267 | 287 | 302 | 334 | 357 | ||
index area time (ms) | 131 | 136 | 155 | 182 | 200 | 219 | 251 | 274 | |||
count(*) | with timeseriesMetadata | overall cost time (ms) | 89 | 90 | 91 | 93 | 93 | 93 | 94 | 97 | 97 |
index area time (ms) | 15 | 16 | 16 | 16 | 16 | 16 | 16 | 17 | 17 | ||
without timeseriesMetadata | overall cost time (ms) | 122 | 123 | 127 | 127 | 127 | 127 | 128 | 130 | ||
index area time (ms) | 50 | 50 | 50 | 50 | 51 | 52 | 52 | 53 |
1000 timeseries in one tsfile: (query for 1 timeseries as well)
chunk number | 1 | 2 | 3 | 5 | 8 | 10 | 15 | 20 | 25 | ||
raw | with timeseriesMetadata | overall cost time (ms) | 421 | 478 | 550 | 673 | 910 | 998 | 1394 | 1637 | 1966 |
index area time (ms) | 274 | 332 | 403 | 528 | 763 | 853 | 1249 | 1496 | 1795 | ||
without timeseriesMetadata | overall cost time (ms) | 489 | 537 | 672 | 903 | 1010 | 1371 | 1650 | 1938 | ||
index area time (ms) | 340 | 393 | 528 | 758 | 864 | 1232 | 1511 | 1789 | |||
count(*) | with timeseriesMetadata | overall cost time (ms) | 260 | 271 | 290 | 331 | 399 | 397 | 562 | 609 | 647 |
index area time (ms) | 133 | 142 | 158 | 197 | 265 | 267 | 427 | 472 | 513 | ||
without timeseriesMetadata | overall cost time (ms) | 307 | 326 | 359 | 428 | 447 | 583 | 620 | 713 | ||
index area time (ms) | 177 | 195 | 227 | 296 | 315 | 447 | 486 | 553 |
Conclusion:
- Although the index area structure with no TimeseriesMetadata speeds up a little in raw data query,
it reduces the speed a lot in aggregation query. => We should reserve TimeseriesMetadata. - The time cost does not change in the data area of TsFile.
(II) Experiment about combine Chunk and Page
Do we need Chunk and Page, or reserve one is ok?
How many points can a chunk have when chunk size = 64K, 1M, 2M, 3M, and 4M?
(1) Write one timeseries in one TsFile, with long data type , random data.
(2) And adjust the number of points by the size of chunk.
try (TsFileWriter tsFileWriter = new TsFileWriter(f)) { // only one timeseries tsFileWriter.registerTimeseries( new Path(Constant.DEVICE_PREFIX, Constant.SENSOR_1), new UnaryMeasurementSchema(Constant.SENSOR_1, TSDataType.INT64, TSEncoding.RLE)); // construct TSRecord for (int i = 1; i <= 7977; i++) { // change here TSRecord tsRecord = new TSRecord(i, Constant.DEVICE_PREFIX); DataPoint dPoint1 = new LongDataPoint(Constant.SENSOR_1, random.nextLong()); tsRecord.addTuple(dPoint1); // write TSRecord tsFileWriter.write(tsRecord); } }
Here are the results:
chunk size | ~64K | ~1M | ~2M | ~3M | ~4M |
points number | 7,977 | 125,000 | 260,000 | 390,000 | 520,000 |
page number | 1 | 16 | 32 | 49 | 66 |
page size (uncompressed) | 65398 | 65398 | 65398 | 65398 | 65398 |
page size (compressed) | 64275 | 64275 | 64275 | 64275 | 64275 |
Discuss the scenarios below: (only one timeseries)
1. For a scenario that generates 5 data points per second. (one chunk one day) (5Hz frequency)
One day will generate 432,000 points (about 54 pages). Therefore, 1 chunk has 54 pages (about 3.4M).
2. For a scenario that generates one data point per second. (one chunk one day) (1Hz frequency)
One day will generate 86,400 points (about 11 pages). Therefore, 1 chunk has 11 pages (about 693K).
3. For a scenario that generates 5 data points per minute. (one chunk one day) (1/12Hz frequency)
One day will generate 7200 points (about 1 pages). Therefore, 1 chunk has 1 page (about 56.6K).
4. For a scenario that generates one data point per minute. (one chunk one week) (1/60Hz frequency)
One week will generate 10080 points (about 1.3 pages). Therefore, 1 chunk has 1~2 pages (about 79.3K).
Reserve both chunk and page:
- Chunk and Page are 2 levels of indexes in one TsFile, Suitable for aggregation and time filter with different granularity.
- Chunk is the unit for I/O and page is the unit for query
- When one Chunk has multiple pages, this structure is better.
Reserve only page:
- one level index in one TsFile.
- Suitable for small Chunk (Mass Timeseries) scenario, in which 1 chunk has only 1~2 pages
(Note: Since 0.12, If one Chunk has only one Page, then PageStatistics will be removed, we only store statistics in ChunkMetadata)
(III) Experiment about how to store PageHeader
(a) store PageHeader with PageData (current design)
(b) combine PageHeader with ChunkHeader
For raw data query in a Chunk:
(1) time > t:
- (a) 顺序读 前几个 Page,然后开始顺序读后面的 PageData
- (b) 顺序读 前几个 PageHeader,然后开始顺序读后面的 PageData
(2) time < t:
- (a) 顺序读前几个符合时间过滤条件的 Page
- (b) 顺序读前几个 PageHeader,然后开始顺序读一部分的 PageData
For aggregation query in a Chunk:
(1) time > t:
- (a) 跳读 所有 PageHeader,获得聚合结果
- (b) 顺序读 所有 PageHeader,获得聚合结果
(2) time < t:
- (a) 跳读 前几个符合时间过滤条件的 PageHeader,获得聚合结果
- (b) 顺序读 前几个符合时间过滤条件的 PageHeader,获得聚合结果
Conclusion:
从理论分析,(b) 方案无论在原始数据查询还是在聚合查询中均会有较好的表现。
使用 (a) 方案仅仅是因为将对应 PageHeader 和 PageData 放在一起存储,易于理解。