mvstore.html 14.2 KB
Newer Older
1 2 3 4 5 6 7 8 9
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<!--
Copyright 2004-2011 H2 Group. Multiple-Licensed under the H2 License, Version 1.0,
and under the Eclipse Public License, Version 1.0
(http://h2database.com/html/license.html).
Initial Developer: H2 Group
-->
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">
<head><meta http-equiv="Content-Type" content="text/html;charset=utf-8" /><title>
10
MVStore
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
</title><link rel="stylesheet" type="text/css" href="stylesheet.css" />
<!-- [search] { -->
<script type="text/javascript" src="navigation.js"></script>
</head><body onload="frameMe();">
<table class="content"><tr class="content"><td class="content"><div class="contentDiv">
<!-- } -->

<h1>MVStore</h1>
<a href="#overview">
    Overview</a><br />
<a href="#example_code">
    Example Code</a><br />
<a href="#features">
    Features</a><br />
<a href="#differences">
    Similar Projects and Differences to Other Storage Engines</a><br />
<a href="#current_state">
    Current State</a><br />
<a href="#building_mvstore">
    Building the MVStore Library</a><br />
<a href="#requirements">
    Requirements</a><br />

<h2 id="overview">Overview</h2>
35 36 37
<p>
The MVStore is work in progress, and is planned to be the next storage subsystem of H2.
</p>
38 39 40 41 42 43
<ul><li>MVStore stands for multi-version store.
</li><li>Each store contains a number of maps (using the <code>java.util.Map</code> interface).
</li><li>The data can be persisted to disk (like a key-value store or a database).
</li><li>Fully in-memory operation is supported.
</li><li>It is intended to be fast, simple to use, and small.
</li><li>Old versions of the data can be read concurrently with all other operations.
44
</li><li>Transaction are supported (currently only one transaction at a time).
45 46 47
</li><li>Transactions (even if they are persisted) can be rolled back.
</li><li>The tool is very modular. It supports pluggable data types / serialization,
pluggable map implementations (B-tree and R-tree currently), BLOB storage,
48
and a file system abstraction to support encryption and compressed read-only files.
49 50 51
</li></ul>

<h2 id="example_code">Example Code</h2>
52 53 54 55 56
<h3>Map Operations and Versioning</h3>
<p>
The following sample code show how to create a store,
open a map, add some data, and access the current and an old version.
</p>
57 58 59 60
<pre>
// open the store (in-memory if fileName is null)
MVStore s = MVStore.open(fileName);

61
// create/get the map named "data"
62
MVMap&lt;Integer, String&gt; map = s.openMap("data");
63 64

// add some data
65 66
map.put(1, "Hello");
map.put(2, "World");
67 68 69 70 71 72 73 74 75 76

// get the current version, for later use
long oldVersion = s.getCurrentVersion();

// from now on, the old version is read-only
s.incrementVersion();

// more changes, in the new version
// changes can be rolled back if required
// changes always go into 'head' (the newest version)
77 78
map.put(1, "Hi");
map.remove(2);
79 80

// access the old data (before incrementVersion)
81
MVMap&lt;Integer, String&gt; oldMap =
82 83 84 85 86 87 88 89
        map.openVersion(oldVersion);

// store the newest data to disk
s.store();

// print the old version (can be done
// concurrently with further modifications)
// this will print Hello World
90 91 92
System.out.println(oldMap.get(1));
System.out.println(oldMap.get(2));
oldMap.close();
93 94

// print the newest version ("Hi")
95
System.out.println(map.get(1));
96 97 98 99 100

// close the store - this doesn't write to disk
s.close();
</pre>

101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
<h3>Store Builder</h3>
<p>
The <code>MVStoreBuilder</code> provides a fluid interface
to build a store if more complex configuration options are used.
</p>
<pre>
MVStore s = MVStoreBuilder.
    fileBased(fileName).
    cacheSizeMB(10).
    readOnly().
    open();
</pre>

<h3>R-Tree</h3>
<p>
The <code>MVRTreeMap</code> is an R-tree implementation
that supports fast spatial queries.
</p>
<pre>
// create an in-memory store
MVStore s = MVStore.open(null);

// create an R-tree map
// the key has 2 dimensions, the value is a string
125
MVRTreeMap&lt;String&gt; r = MVRTreeMap.create(2, new ObjectType());
126 127 128 129 130 131 132 133 134 135 136

// open the map
r = s.openMap("data", r);

// add two key-value pairs
// the first value is the key id (to make the key unique)
// then the min x, max x, min y, max y
r.add(new SpatialKey(0, -3f, -2f, 2f, 3f), "left");
r.add(new SpatialKey(1, 3f, 4f, 4f, 5f), "right");

// iterate over the intersecting keys
137
Iterator&lt;SpatialKey&gt; it =
138 139 140 141 142 143 144 145
        r.findIntersectingKeys(new SpatialKey(0, 0f, 9f, 3f, 6f));
for (SpatialKey k; it.hasNext();) {
    k = it.next();
    System.out.println(k + ": " + r.get(k));
}
s.close();
</pre>

146 147 148 149 150 151

<h2 id="features">Features</h2>

<h3>Maps</h3>
<p>
Each store supports a set of named maps.
152
A map is sorted by key, and supports the common lookup operations,
153 154
including access to the first and last key, iterate over some or all keys, and so on.
</p><p>
155 156 157 158 159 160 161
Also supported, and very uncommon for maps, is fast index lookup.
The keys of the map can be accessed like a list
(get the key at the given index, get the index of a certain key).
That means getting the median of two keys is trivial,
and it allows to very quickly count ranges.
The iterator supports fast skipping.
This is possible because internally, each map is organized in the form of a counted B+-tree.
162
</p><p>
163
In database terms, a map can be used like a table, where the key of the map is the primary key of the table,
164
and the value is the row. A map can also represent an index, where the key of the map is the key
165 166
of the index, and the value of the map is the primary key of the table (for non-unique indexes
the key of the map must also contain the primary key).
167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194
</p>

<h3>Versions / Transactions</h3>
<p>
Multiple versions are supported.
A version is a snapshot of all the data of all maps at a given point in time.
A transaction is a number of actions between two versions.
</p><p>
Versions / transactions are not immediately persisted; instead, only the version counter is incremented.
If there is a change after switching to a new version, a snapshot of the old version is kept in memory,
so that the old version can still be read.
</p><p>
Old persisted versions are readable until the old data was explicitly overwritten.
Creating the snapshot is fast: only the pages that are changed after a snapshot are copied.
This behavior also called COW (copy on write).
</p><p>
Rollback is supported (rollback to any old in-memory version or an old persisted version).
</p>

<h3>Log Structured Storage</h3>
<p>
Currently, store() needs to be called explicitly to save changes.
Changes are buffered in memory, and once enough changes have accumulated
(for example 2 MB), all changes are written in one continuous disk write operation.
But of course, if needed, changes can also be persisted if only little data was changed.
The estimated amount of unsaved changes is tracked.
The plan is to automatically store in a background thread once there are enough changes.
</p><p>
195 196 197
When storing, all changed pages are serialized,
compressed using the LZF algorithm (this can be disabled),
and written sequentially to a free area of the file.
198
Each such change set is called a chunk.
199 200
All parent pages of the changed B-trees are stored in this chunk as well,
so that each chunk also contains the root of each changed map
201 202 203 204
(which is the entry point to read old data).
There is no separate index: all data is stored as a list of pages.
</p><p>
There are currently two write operations per chunk:
205 206
one to store the chunk data (the pages), and one to update the file header
(so it points to the chunk head), but the plan is to write the file header only
207 208
once in a while, so it does not slow down opening the store too much.
</p><p>
209
There is currently no transaction log, no undo log,
210 211 212 213 214
and there are no in-place updates (however unused chunks are overwritten).
To efficiently persist very small transactions, the plan is to support a transaction log
where only the deltas is stored, until enough changes have accumulated to persist a chunk.
Old versions are kept and are readable until they are no longer needed.
</p><p>
215
The plan is to keep all old data for at least one or two minutes (configurable),
216 217 218 219 220 221
so that there are no explicit sync operations required to guarantee data consistency.
To reuse disk space, the chunks with the lowest amount of live data are compacted
(the live data is simply stored again in the next chunk).
To improve data locality and disk space usage, the plan is to automatically defragment and compact data.
</p><p>
Compared to regular databases that use a transaction log, undo log, and main storage area,
222
the log structured storage is simpler, more flexible, and typically needs less disk operations per change,
223 224 225 226 227 228 229 230 231 232 233 234
as data is only written once instead of twice or 3 times, and because the B-tree pages are
always full (they are stored next to each other) and can be easily compressed.
But temporarily, disk space usage might actually be a bit higher than for a regular database,
as disk space is not immediately re-used (there are no in-place updates).
</p>

<h3>In-Memory Performance and Usage</h3>
<p>
Performance of in-memory operations is comparable with <code>java.util.TreeMap</code>
(many operations are actually faster), but usually slower than <code>java.util.HashMap</code>.
</p><p>
If no file name is specified, the store operates purely in memory.
235
Except for persisting data, all features are supported in this mode
236 237 238
(multi-versioning, index lookup, R-tree and so on).
</p><p>
The memory overhead for large maps is slightly better than for the regular
239 240
map implementations, but there is a higher overhead per map.
For maps with less than 25 entries, the regular map implementations
241 242 243 244 245
use less memory on average.
</p>

<h3>Pluggable Data Types</h3>
<p>
246
Serialization is pluggable. The default serialization currently supports many common data types,
247 248 249 250
and uses Java serialization for other objects. The following classes are currently directly supported:
<code>Boolean, Byte, Short, Character, Integer, Long, Float, Double, BigInteger, BigDecimal,
byte[], char[], int[], long[], String, UUID</code>.
The plan is to add more common classes (date, time, timestamp, object array).
251 252 253 254
</p><p>
Parameterized data types are supported
(for example one could build a string data type that limits the length for some reason).
</p><p>
255
The storage engine itself does not have any length limits, so that keys, values,
256 257 258 259 260 261 262
pages, and chunks can be very big (as big as fits in memory).
Also, there is no inherent limit to the number of maps and chunks.
Due to using a log structured storage, there is no special case handling for large keys or pages.
</p>

<h3>BLOB Support</h3>
<p>
263
There is a mechanism that stores large binary objects by splitting them into smaller blocks.
264 265 266 267 268 269 270 271
This allows to store objects that don't fit in memory.
Streaming as well as random access reads on such objects are supported.
This tool is written on top of the store (only using the map interface).
</p>

<h3>R-Tree and Pluggable Map Implementations</h3>
<p>
The map implementation is pluggable.
272
In addition to the default MVMap (multi-version map),
273 274 275 276 277 278 279
there is a multi-version R-tree map implementation
for spatial operations (contain and intersection; nearest neighbor is not yet implemented).
</p>

<h3>Concurrent Operations and Caching</h3>
<p>
At the moment, concurrent read on old versions of the data is supported.
280
All such read operations can occur in parallel. Concurrent reads from the page cache,
281 282 283
as well as concurrent reads from the file system are supported.
</p><p>
Caching is done on the page level.
284
The page cache is a concurrent LIRS cache,
285 286
which should be resistant against scan operations.
</p><p>
287 288
Concurrent modification operations on the maps are currently not supported,
however it is planned to support an additional map implementation
289
that supports concurrent writes
290
(at the cost of speed if used in a single thread, same as <code>ConcurrentHashMap</code>).
291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
</p>

<h3>File System Abstraction, File Locking and Online Backup</h3>
<p>
The file system is pluggable (the same file system abstraction is used as H2 uses).
Support for encryption is planned using an encrypting file system.
Other file system implementations support reading from a compressed zip or tar file.
</p>
<p>
Each store may only be opened once within a JVM.
When opening a store, the file is locked in exclusive mode, so that
the file can only be changed from within one process.
Files can be opened in read-only mode, in which case a shared lock is used.
</p>
<p>
The persisted data can be backed up to a different file at any time,
307
even during write operations (online backup).
308
To do that, automatic disk space reuse needs to be first disabled, so that
309
new data is always appended at the end of the file.
310
Then, the file can be copied (the file handle is available to the application).
311 312
</p>

313 314
<h3>Tools</h3>
<p>
315
There is a builder for store instances (<code>MVStoreBuilder</code>)
316 317 318 319 320
with a fluent API to simplify building a store instance.
</p>
<p>
There is a tool (<code>MVStoreTool</code>) to dump the contents of a file.
</p>
321 322 323 324 325 326

<h2 id="differences">Similar Projects and Differences to Other Storage Engines</h2>
<p>
Unlike similar storage engines like LevelDB and Kyoto Cabinet, the MVStore is written in Java
and can easily be embedded in a Java application.
</p><p>
327
The MVStore is somewhat similar to the Berkeley DB Java Edition because it is also written in Java,
328
and is also a log structured storage, but the H2 license is more liberal.
329 330 331 332 333 334
</p><p>
Like SQLite, the MVStore keeps all data in one file.
The plan is to make the MVStore easier to use and faster than SQLite on Android
(this was not recently tested, however an initial test was successful).
</p><p>
The API of the MVStore is similar to MapDB (previously known as JDBM) from Jan Kotek,
335
and some code is shared between MapDB and JDBM.
336
However, unlike MapDB, the MVStore uses is a log structured storage.
337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
</p>

<h2 id="current_state">Current State</h2>
<p>
The code is still very experimental at this stage.
The API as well as the behavior will probably change.
Features may be added and removed (even thought the main features will stay).
</p>

<h2 id="building_mvstore">Building the MVStore Library</h2>
<p>
There is currently no build script.
To test it, run the test within the H2 project in Eclipse or any other IDE.
</p>

<h2 id="requirements">Requirements</h2>
<p>
There are no special requirements.
The MVStore should run on any JVM as well as on Android
(even thought this was not tested recently).
</p>

<!-- [close] { --></div></td></tr></table><!-- } --><!-- analytics --></body></html>