panthema / 2007 / stx-btree
Design of a small B+ tree

STX B+ Tree C++ Template Classes

Gravatar Eduard: a) also ich verstehe diesen "Sicherheits"-Einwand auch nicht

b) wäre es möglich, den Code unter einer anderen Lizenz zu stellen? LGPL ist eigentlich nur sinnvoll für shared-Libs, bei template-Headers wirkt es ähnlich restriktiv wie die GPL. BSD-ähnliche Lizenzen sind IMHO im allgemeinen brauchbarer, sogar die alte BSD-Lizenz mit Advertising-Klausel. Siehe auch GCC-STL-Headers, stehen unter BSD-artigen SGI-Lizenz oder LGPL aber mit deutlichen Ausnahmen.
Gravatar Timo: Ein paar Worte zu deinen drei Punkten: Ja, die Implementierung macht natürlich eine binäre Suche auf den sortierten Schlüsselwerten in einem Knoten. Jedoch würde ich binäre Bäume nicht als Sicherheitsrisiko, sondern als Performance- oder Skalierungsproblem bezeichnen. Der B-Baum verbraucht verglichen mit dem Binärbaum zwar insgesamt mehr Speicher, aber immer in zusammenhängenden Blöcken. Ob virtuellen oder realen Speicher ist erstmal egal. Durch die größeren Speicherblöcke wird die L1/L2 Cache-Hitrate verbessert, dies zu zeigen war eines der Ausgangspunkte, wie auch in der Zusammenfassung steht. Was an dem direkten Vergleich "unfair" sein soll verstehe ich nicht. Er war Ausgangsüberlegung der Arbeit. Darüber steht aber viel mehr im englischen Text.
Gravatar Frank Mertens: Der Performance-Vergleich ist etwas unfair gegenüber den Binär-Bäumen. Wäre schön, wenn folgende Punkte Erwähnung fänden:
  • Eine B-Baum Implementierung sollte binäre Suche auf den Knoten verwenden. (Macht STX das?)
  • Binäre Bäume sind ein Sicherheitsrisiko, da die Knoten beliebig im Speicher gestreut sein können. Das kann zu Seitenflattern führen. B-Bäume skalieren dagenen sehr viel besser im virtuellem Address-Raum.
  • Die Knoten der B-Bäume clustern Elemente im Speicher und führen dadruch zu besserer L1/L2-Cache-Nutzung. (Daher vermutlich ab 16000 Integer etwas schneller...)

Posted on 2007-04-27, last updated 2013-05-05 by Timo Bingmann at Permlink.

Thumbnail of the B+ demo program
demo download page


The STX B+ Tree package is a set of C++ template classes implementing a B+ tree key/data container in main memory. The classes are designed as drop-in replacements of the STL containers set, map, multiset and multimap and follow their interfaces very closely. By packing multiple value pairs into each node of the tree the B+ tree reduces heap fragmentation and utilizes cache-line effects better than the standard red-black binary tree. The tree algorithms are based on the implementation in Cormen, Leiserson and Rivest's Introduction into Algorithms, Jan Jannink's paper and other algorithm resources. The classes contain extensive assertion and verification mechanisms to ensure the implementation's correctness by testing the tree invariants. To illustrate the B+ tree's structure a wxWidgets demo program is included in the source package.

The main B+ tree implementation can be found in doxygen stx/btree.h or with plain text comments btree.h.

Special interest was put into performing a speed comparison test between the standard red-black tree and the new B+ tree implementation. The speed test results are interesting and show the B+ tree to be significantly faster for trees containing more than 16,000 items on a Pentium 4 and more than 100,000 items on an Intel Core i7 CPU.

The B+ tree main header code is covered to 90.3% by test cases. A graphical display of the test suite's coverage can be viewed online.

The package includes a demo program, which illustrates how the B+ tree organises integer and string keys. Compiled binaries for Windows and and some Linux distributions are available on the demo download page.

See the README file below for a more detailed overview. See ChangeLog below on what changed in version 0.9.


STX B+ Tree Version 0.9 (current) released 2013-05-05
Source code archive:
(includes Doxygen HTML)
stx-btree-0.9.tar.bz2 stx-btree-0.9.tar.bz2 (1.65 MiB) Browse online
Extensive Documentation: Browse documentation online
Demo Binaries: See Extra Download Page for Win32 and Linux binaries.

See bottom of this page for older downloads.

License and Git Repository

The B+ tree template source code is released under the Boost Software License, Version 1.0, which can be found at the header of each include file.

All auxiliary programs like the wxWidgets demo, test suite and speed tests are licensed under the GNU General Public License v3 (GPLv3).

The git repository containing all sources and packages is available by
git clone
Some further papers, documentation and some future branches are also available there.


Original Idea

The idea originally arose while coding a read-only database, which used a huge map of millions of non-sequential integer keys to 8-byte file offsets. When using the standard STL red-black tree implementation this would yield millions of 20-byte heap allocations and very slow search times due to the tree's height. So the original intension was to reduce memory fragmentation and improve search times. The B+ tree solves this by packing multiple data pairs into one node with a large number of descendant nodes.

In computer science lectures it is often stated that using consecutive bytes in memory would be more cache-efficient, because the CPU's cache levels always fetch larger blocks from main memory. So it would be best to store the keys of a node in one continuous array. This way the inner scanning loop would be accelerated by benefiting from cache effects and pipelining speed-ups. Thus the cost of scanning for a matching key would be lower than in a red-black tree, even though the number of key comparisons are theoretically larger. This second aspect aroused my academic interest and resulted in the speed test experiments.

A third inspiration was that no working C++ template implementation of a B+ tree could be found on the Internet. Now this one can be found.

Implementation Overview

This implementation contains five main classes within the stx namespace (blandly named Some Template eXtensions). The base class btree implements the B+ tree algorithms using inner and leaf nodes in main memory. Almost all STL-required function calls are implemented (see below for the exceptions). The asymptotic time requirements of the STL standard are theoretically not always fulfilled. However in practice this B+ tree performs better than the STL's red-black tree at the cost of using more memory. See the speed test results for details.

The base class is then specialized into btree_set, btree_multiset, btree_map and btree_multimap using default template parameters and facade functions. These classes are designed to be drop-in replacements for the corresponding STL containers.

The insertion function splits the nodes on recursion unroll. Erase is largely based on Jannink's ideas. See for his paper on "Implementing Deletion in B+-trees".

The two set classes (btree_set and btree_multiset) are derived from the base implementation class btree by specifying an empty struct as data_type. All functions are adapted to provide the base class with empty placeholder objects. Note that it is somewhat inefficient to implement a set or multiset using a B+ tree: a plain B tree (without +) would hold no extra copies of the keys. The main focus was on implementing the maps.

Problem with Separated Key/Data Arrays

The most noteworthy difference to the default red-black tree implementation of std::map is that the B+ tree does not hold key/data pairs together in memory. Instead each B+ tree node has two separate arrays containing keys and data values. This design was chosen to utilize cache-line effects while scanning the key array.

However it also directly generates many problems in implementing the iterators' operators. These return a (writable) reference or pointer to a value_type, which is a std::pair composition. These data/key pairs however are not stored together and thus a temporary copy must be constructed. This copy should not be written to, because it is not stored back into the B+ tree. This effectively prohibits use of many STL algorithms which writing to the B+ tree's iterators. I would be grateful for hints on how to resolve this problem without folding the key and data arrays.

Test Suite

The B+ tree distribution contains an extensive test suite using cppunit. According to gcov 90.9% of the btree.h implementation is covered.

STL Incompatibilities

Key and Data Type Requirements

The tree algorithms currently do not use copy-construction. All key/data items are allocated in the nodes using the default-constructor and are subsequently only assigned new data (using operator=).

Iterators' Operators

The most important incompatibility are the non-writable operator* and operator-> of the iterator. See above for a discussion of the problem on separated key/data arrays. Instead of *iter and iter-> use the new function which returns a writable reference to the data value in the tree.

Erase Functions

The B+ tree supports three functions:

size_type erase(const key_type &key); // erase all data pairs matching key
bool erase_one(const key_type &key);  // erase one data pair matching key
void erase(iterator iter);            // erase pair referenced by iter

The following STL-required function is not supported:

void erase(iterator first, iterator last);


Beyond the usual STL interface the B+ tree classes support some extra goodies.

// Bulk load a sorted range. Loads items into leaves and constructs a
// B-tree above them. The tree must be empty when calling this function.
template <typename Iterator>
void bulk_load(Iterator ibegin, Iterator iend);

// Output the tree in a pseudo-hierarchical text dump to std::cout. This
// function requires that BTREE_DEBUG is defined prior to including the btree
// headers. Furthermore the key and data types must be std::ostream printable.
void print() const;

// Run extensive checks of the tree invariants. If a corruption in found the
// program will abort via assert(). See below on enabling auto-verification.
void verify() const;

// Serialize and restore the B+ tree nodes and data into/from a binary image.
// This requires that the key and data types are integral and contain no
// outside pointers or references.
void dump(std::ostream &os) const;
bool restore(std::istream &is);

B+ Tree Traits

All tree template classes take a template parameter structure which holds important options of the implementation. The following structure shows which static variables specify the options and the corresponding defaults:

struct btree_default_map_traits
    // If true, the tree will self verify it's invariants after each insert()
    // or erase(). The header must have been compiled with BTREE_DEBUG
    // defined.
    static const bool   selfverify = false;

    // If true, the tree will print out debug information and a tree dump
    // during insert() or erase() operation. The header must have been
    // compiled with BTREE_DEBUG defined and key_type must be std::ostream
    // printable.
    static const bool   debug = false;

    // Number of slots in each leaf of the tree. Estimated so that each node
    // has a size of about 256 bytes.
    static const int    leafslots =
                             MAX( 8, 256 / (sizeof(_Key) + sizeof(_Data)) );

    // Number of slots in each inner node of the tree. Estimated so that each
    // node has a size of about 256 bytes.
    static const int    innerslots =
                             MAX( 8, 256 / (sizeof(_Key) + sizeof(void*)) );

    // As of stx-btree-0.9, the code does linear search in find_lower() and
    // find_upper() instead of binary_search, unless the node size is larger
    // than this threshold. See notes at
    static const size_t binsearch_threshold = 256;

Speed Tests

See the web page for speed test results and a discussion thereof.


2013-05-05 - Timo Bingmann - v0.92011-05-17 - Timo Bingmann - v0.8.6 2011-05-06 - Timo Bingmann - v0.8.6 2011-05-03 - Timo Bingmann - v0.8.6 2008-09-07 - Timo Bingmann - v0.8.3 2008-09-03 - Timo Bingmann - v0.8.3 2008-08-13 - Timo Bingmann - v0.8.2 2008-08-01 - Timo Bingmann - v0.8.2 2008-08-01 - Timo Bingmann - v0.8.2 2008-01-25 - Timo Bingmann - v0.8.1 2007-05-12 - Timo Bingmann - v0.8

Older Downloads

STX B+ Tree Version 0.8.6 released 2011-05-18
Source code archive:
(includes Doxygen HTML)
stx-btree-0.8.6.tar.bz2 stx-btree-0.8.6.tar.bz2 (1.70 MiB)
MD5: 552ca8419ce21b75af2fbb74aea4e253
Browse online
Extensive Documentation: Browse documentation online

STX B+ Tree Version 0.8.3 released 2008-09-07
Source code archive:
(includes Doxygen HTML)
stx-btree-0.8.3.tar.bz2 stx-btree-0.8.3.tar.bz2 (931 KiB)
MD5: 1c13439c5d6ca6ba8bfce6b39f1ca65c
Browse online
Extensive Documentation: Browse documentation online

STX B+ Tree Version 0.8.2 released 2008-08-13
Source code archive:
(includes Doxygen HTML)
stx-btree-0.8.2.tar.bz2 stx-btree-0.8.2.tar.bz2 (787 KiB)
MD5: bf147a1f2f9a540d283244e5a92c5353
Browse online
Extensive Documentation: Browse documentation online

STX B+ Tree Version 0.8.1 released 2008-01-25
Source code archive: stx-btree-0.8.1.tar.bz2 stx-btree-0.8.1.tar.bz2 (411 KiB)
MD5: 87df74dab5c5b2a34c6ebfbfc224b26b
Browse online
Extensive Documentation: stx-btree-0.8.1-doxygen.tar.bz2 stx-btree-0.8.1-doxygen.tar.bz2 (309 KiB)
MD5: f7801dd6e8672820a599704a7fd7df4f
Browse documentation online

STX B+ Tree Version 0.8 released 2007-05-13
Source code archive: stx-btree-0.8.tar.bz2 stx-btree-0.8.tar.bz2 (411 KiB)
MD5: b3e2981dff63d9a01bfc0a102a49c32c
Browse online
Extensive Documentation: stx-btree-0.8-doxygen.tar.bz2 stx-btree-0.8-doxygen.tar.bz2 (324 KiB)
MD5: 7e14e8eb904129f77d96c8abb517068d
Browse documentation online

STX B+ Tree Version 0.7 released 2007-04-27
Source code archive: stx-btree-0.7.tar.bz2 stx-btree-0.7.tar.bz2 (359 KiB)
MD5: b10da911facd14f4faa6f31b43fd0591
Browse online
Extensive Documentation: stx-btree-0.7-doxygen.tar.bz2 stx-btree-0.7-doxygen.tar.bz2 (291 KiB)
MD5: a4106a81fb5982a3bc5fcb822f85d219
Browse documentation online
RSS 2.0 Weblog Feed Atom 1.0 Weblog Feed Valid XHTML 1.1 Valid CSS (2.1)
Copyright 2005-2014 Timo Bingmann - Impressum