4 <title>Sixth Data - Data storage and computing engine</title>
5 <!-- 2019-09-20 Fri 09:53 -->
7 <meta name="viewport" content="width=device-width, initial-scale=1">
8 <meta name="generator" content="Org-mode">
9 <meta name="author" content="Svjatoslav Agejenko">
10 <link href="https://bootswatch.com/3/darkly/bootstrap.min.css" rel="stylesheet">
11 <script src="https://cdnjs.cloudflare.com/ajax/libs/jquery/1.11.3/jquery.min.js"></script>
12 <script src="https://cdnjs.cloudflare.com/ajax/libs/twitter-bootstrap/3.3.5/js/bootstrap.min.js"></script>
13 <style type="text/css">
14 footer {background-color: #111 !important;}
15 pre {background-color: #111; color: #ccc;}
17 <style type="text/css">
18 /* org mode styles on top of twbs */
35 background-color: #f5f5f5;
62 display: inline-block;
74 .bs-docs-sidebar.affix {
77 @media (min-width: 768px) {
83 /* All levels of nav */
84 .bs-docs-sidebar .nav > li > a {
91 .bs-docs-sidebar .nav > li > a:hover,
92 .bs-docs-sidebar .nav > li > a:focus {
95 text-decoration: none;
96 background-color: transparent;
97 border-left: 1px solid #A1283B;
99 .bs-docs-sidebar .nav > .active > a,
100 .bs-docs-sidebar .nav > .active:hover > a,
101 .bs-docs-sidebar .nav > .active:focus > a {
105 background-color: transparent;
106 border-left: 2px solid #A1283B;
109 /* Nav: second level (shown on .active) */
110 .bs-docs-sidebar .nav .nav {
111 display: none; /* Hide by default, but at >768px, show it */
112 padding-bottom: 10px;
114 .bs-docs-sidebar .nav .nav > li > a {
121 .bs-docs-sidebar .nav .nav > li > a:hover,
122 .bs-docs-sidebar .nav .nav > li > a:focus {
125 .bs-docs-sidebar .nav .nav > .active > a,
126 .bs-docs-sidebar .nav .nav > .active:hover > a,
127 .bs-docs-sidebar .nav .nav > .active:focus > a {
132 /* Nav: third level (shown on .active) */
133 .bs-docs-sidebar .nav .nav .nav {
134 padding-bottom: 10px;
136 .bs-docs-sidebar .nav .nav .nav > li > a {
143 .bs-docs-sidebar .nav .nav .nav > li > a:hover,
144 .bs-docs-sidebar .nav .nav .nav > li > a:focus {
147 .bs-docs-sidebar .nav .nav .nav > .active > a,
148 .bs-docs-sidebar .nav .nav .nav > .active:hover > a,
149 .bs-docs-sidebar .nav .nav .nav > .active:focus > a {
154 /* Show and affix the side nav when space allows it */
155 @media (min-width: 992px) {
156 .bs-docs-sidebar .nav > .active > ul {
159 /* Widen the fixed sidebar */
160 .bs-docs-sidebar.affix,
161 .bs-docs-sidebar.affix-bottom {
164 .bs-docs-sidebar.affix {
165 position: fixed; /* Undo the static from mobile first approach */
168 .bs-docs-sidebar.affix-bottom {
169 position: absolute; /* Undo the static from mobile first approach */
171 .bs-docs-sidebar.affix .bs-docs-sidenav,.bs-docs-sidebar.affix-bottom .bs-docs-sidenav {
176 @media (min-width: 1200px) {
177 /* Widen the fixed sidebar again */
178 .bs-docs-sidebar.affix-bottom,
179 .bs-docs-sidebar.affix {
184 <script type="text/javascript">
188 $('.bs-docs-sidebar li').first().addClass('active');
190 $(document.body).scrollspy({target: '.bs-docs-sidebar'});
192 $('.bs-docs-sidebar').affix();
197 <div id="content" class="container">
198 <div class="row"><div class="col-md-9"><h1 class="title">Sixth Data - Data storage and computing engine</h1>
200 <div id="outline-container-sec-1" class="outline-2">
201 <h2 id="sec-1"><span class="section-number-2">1</span> General</h2>
202 <div class="outline-text-2" id="text-1">
204 <li>This is a subproject of <a href="https://www3.svjatoslav.eu/projects/sixth/">Sixth</a>
207 <li>This program is free software: you can redistribute it and/or modify
208 it under the terms of the <a href="https://www.gnu.org/licenses/lgpl.html">GNU Lesser General Public License</a> as
209 published by the Free Software Foundation, either version 3 of the
210 License, or (at your option) any later version.
215 <li>Svjatoslav Agejenko
217 <li>Homepage: <a href="https://svjatoslav.eu">https://svjatoslav.eu</a>
219 <li>Email: <a href="mailto://svjatoslav@svjatoslav.eu">mailto://svjatoslav@svjatoslav.eu</a>
224 <li><a href="https://www.svjatoslav.eu/projects/">Other software projects hosted at svjatoslav.eu</a>
229 <div id="outline-container-sec-1-1" class="outline-3">
230 <h3 id="sec-1-1"><span class="section-number-3">1.1</span> Source code</h3>
231 <div class="outline-text-3" id="text-1-1">
233 <li><a href="https://www2.svjatoslav.eu/gitweb/?p=sixth-data.git;a=snapshot;h=HEAD;sf=tgz">Download latest snapshot in TAR GZ format</a>
236 <li><a href="https://www2.svjatoslav.eu/gitweb/?p=sixth-data.git;a=summary">Browse Git repository online</a>
239 <li>Clone Git repository using command:
240 <pre class="example">
241 git clone https://www2.svjatoslav.eu/git/sixth-data.git
246 <li>See <a href="https://www3.svjatoslav.eu/projects/sixth-data/apidocs/">JavaDoc</a>.
253 <div id="outline-container-sec-2" class="outline-2">
254 <h2 id="sec-2"><a id="ID-f6764282-a6f6-44e6-8716-b428074dd093" name="ID-f6764282-a6f6-44e6-8716-b428074dd093"></a><span class="section-number-2">2</span> Vision / goal</h2>
255 <div class="outline-text-2" id="text-2">
257 Provide versioned, clustered, flexible, distributed, multi-dimensional
258 data storage engine for the <a href="http://www2.svjatoslav.eu/gitbrowse/sixth/doc/index.html">Sixth computation engine</a>.
262 <li>Speaking of traditional relational database and object oriented
263 business applications:
266 <li>I hate object-relational impedance mismatch.
269 <li>I don't like to convert data between persistent database and
270 runtime objects for every transaction. How about creating united
271 database/computation engine instead to:
274 <li>Eliminate constant moving and converting of data between 2 systems
275 and make computing happen close to where the data is stored.
278 <li>Abstract away difference between RAM VS persistent storage. Let
279 the system decide at runtime which data to keep in what kind of
288 <div id="outline-container-sec-3" class="outline-2">
289 <h2 id="sec-3"><span class="section-number-2">3</span> Inspiration</h2>
290 <div class="outline-text-2" id="text-3">
292 <li>Relational databases:
296 <li>Indexable / Quickly searchable.
301 <li>Git (version control system)
305 <li>Branchable / mergeable.
307 <li>Transparent cansistency, checksumming and deduplication.
309 <li>(Git as a database:
313 <a href="https://www.kenneth-truyers.net/2016/10/13/git-nosql-database/">https://www.kenneth-truyers.net/2016/10/13/git-nosql-database/</a> )
319 <div id="outline-container-sec-3-1" class="outline-3">
320 <h3 id="sec-3-1"><a id="ID-d2375acc-af14-4f18-8ad0-7949501178c5" name="ID-d2375acc-af14-4f18-8ad0-7949501178c5"></a><span class="section-number-3">3.1</span> Brain</h3>
321 <div class="outline-text-3" id="text-3-1">
323 <li>Brain appears to have more than 3D dimensional design:
324 <a href="https://singularityhub.com/2017/06/21/is-there-a-multidimensional-mathematical-world-hidden-in-the-brains-computation/">https://singularityhub.com/2017/06/21/is-there-a-multidimensional-mathematical-world-hidden-in-the-brains-computation/</a>
327 <li>Brain appears to use geometry to map thoughts and even sounds:
329 <li><a href="https://www.quantamagazine.org/the-brain-maps-out-ideas-and-memories-like-spaces-20190114/">https://www.quantamagazine.org/the-brain-maps-out-ideas-and-memories-like-spaces-20190114/</a>
331 <li><a href="https://www.quantamagazine.org/goals-and-rewards-redraw-the-brains-map-of-the-world-20190328">https://www.quantamagazine.org/goals-and-rewards-redraw-the-brains-map-of-the-world-20190328</a>
336 <li>It directly inspires <a href="#sec-4-1">Geometrical computation</a> idea and nicely fits
337 with <a href="#sec-3-2">CM-1 Connection Machine</a> design.
343 <div id="outline-container-sec-3-2" class="outline-3">
344 <h3 id="sec-3-2"><a id="ID-01aa65c1-3d44-44a8-9b90-58454bc6be80" name="ID-01aa65c1-3d44-44a8-9b90-58454bc6be80"></a><span class="section-number-3">3.2</span> CM-1 Connection Machine</h3>
345 <div class="outline-text-3" id="text-3-2">
347 <a href="https://en.wikipedia.org/wiki/Connection_Machine">https://en.wikipedia.org/wiki/Connection_Machine</a>
351 <li>see: <a href="#sec-4-1">Geometrical computation</a>
354 <li>Computation unit has local CPU and RAM.
357 <li>Data is pre-distributed across computation units.
360 <li>Machine's internal 12-dimensional hypercube network allows to
361 efficiently simulate arbitrary dimensional network topology between
362 computational units. So that when we are solving/simulating for
363 example 5 dimensional problem, we can arrange computational units
364 into virtual 5D network.
371 <div id="outline-container-sec-4" class="outline-2">
372 <h2 id="sec-4"><span class="section-number-2">4</span> Ideas</h2>
373 <div class="outline-text-2" id="text-4">
374 </div><div id="outline-container-sec-4-1" class="outline-3">
375 <h3 id="sec-4-1"><a id="ID-171fe375-c737-41e6-b429-a414f6abc5d8" name="ID-171fe375-c737-41e6-b429-a414f6abc5d8"></a><span class="section-number-3">4.1</span> Geometrical computation</h3>
376 <div class="outline-text-3" id="text-4-1">
378 <li>Inspired by <a href="#sec-3-1">Brain</a>.
380 <li>Wits nicely with <a href="#sec-3-2">CM-1 Connection Machine</a> properties.
385 <div id="outline-container-sec-4-1-1" class="outline-4">
386 <h4 id="sec-4-1-1"><a id="ID-5d287158-53ea-44a2-a754-dd862366066a" name="ID-5d287158-53ea-44a2-a754-dd862366066a"></a><span class="section-number-4">4.1.1</span> Distributed computation and data storage</h4>
387 <div class="outline-text-4" id="text-4-1-1">
389 Maybe every problem can be translated to geometry (use any shapes and
390 as many dimensions as you need). Solution(s) to such problems would
391 then appear as relatively simple search/comparison/lookup results. As
392 a bonus, such geometrical *data storage* AND *computation* can be
393 naturally made in *parallel* and *distributed*. That's what neurons in
394 the brain appear to be doing ! :) . Learning means building/updating
395 the model (the hard part). Question answering is making (relatively
396 simple) lookups (geometrical queries) against the model.
400 <div id="outline-container-sec-4-1-2" class="outline-4">
401 <h4 id="sec-4-1-2"><a id="ID-a117c11e-97c1-4822-88b2-9fc10f96caec" name="ID-a117c11e-97c1-4822-88b2-9fc10f96caec"></a><span class="section-number-4">4.1.2</span> Mapping of hyperspace to traditional object-oriented model</h4>
402 <div class="outline-text-4" id="text-4-1-2">
404 Object oriented programming is inspired by the way human mind
405 operates. It allows programmer to express ideas to computer in a more
410 It is possible to map object model to geometrical hyperspace:
414 <li>Object is a point in space (universe). Each object member variable
415 translates to its own dimension. That is: if class declares 4
416 variables for an object, then corresponding object can be stored as
417 a single point inside 4 dimensional space. Variable values translate
418 to point coordinates in space. That is: Integer, floating point
419 number and even boolean and string can be translated to linear value
420 that can be used as a coordinate along particular dimension.
423 <li>Each class declares its own space (universe). All class instances
424 (objects) are points inside that particular universe. References
425 between objects of different types are hyperlinks (portals) between
431 <div id="outline-container-sec-4-1-3" class="outline-4">
432 <h4 id="sec-4-1-3"><a id="ID-b6b15bd2-c78b-4c51-a343-72843a515c29" name="ID-b6b15bd2-c78b-4c51-a343-72843a515c29"></a><span class="section-number-4">4.1.3</span> Handling of relations</h4>
433 <div class="outline-text-4" id="text-4-1-3">
435 Consider we want to create database of books and authors. Book can
436 have multiple authors, and single person can be author for multiple
437 books. It is possible to store how many hours of work each author has
438 contributed to every book, using hyperspace as follows:
442 <li>Every dimension corresponds to one particular book author. (10
443 authors in the database, would require 10 dimensional space)
445 <li>Point in space corresponds to one particular book.
447 <li>Point location along particular (author) dimension corresponds
448 to amount of work contributed by particular author for given
462 <li>Every dimension corresponds to one particular book.
464 <li>Point in space corresponds to one particular author in the entire
467 <li>Point location along particular (book) dimension corresponds to
468 amount of work contributed for book by given author (point).
479 <div id="outline-container-sec-4-2" class="outline-3">
480 <h3 id="sec-4-2"><span class="section-number-3">4.2</span> Layered architecture</h3>
481 <div class="outline-text-3" id="text-4-2">
483 <dt> layer 1 </dt><dd>disk / block storage / partition
486 <dt> layer 2 </dt><dd>key/value storage. Keys are unique and are dictated by
487 storage engine. Value is arbitrary but limited size byte
488 array. This layer is responsible for handling disk
489 defragmentation and consistency in case of crash
493 <dt> layer 3 </dt><dd>key/value storage. Keys are content hashes. Values are
494 arbitrary but limited size content byte arrays. This
495 layer effectively implements content addressable
496 storage. Content addressible storage enables GIT-like
497 behavior (possibility for competing branches, retaining
498 history, transparent deduplication)
501 <dt> layer 4 </dt><dd>Implements arbitrary dimensional multiverse.
504 <dt> layer 5 </dt><dd>Distributed computation engine.
510 <div id="outline-container-sec-5" class="outline-2">
511 <h2 id="sec-5"><span class="section-number-2">5</span> Current status</h2>
512 <div class="outline-text-2" id="text-5">
514 <li>More or less defined <a href="#sec-2">Vision / goal</a>.
517 <li>Collected some <a href="#sec-3-1">ideas</a>.
520 <li>Implemented very simple persistent key-value map.
522 <li>Long term goal is to use it as a backing storage engine and
523 implement more advanced features on top of this.
531 <div id="outline-container-sec-6" class="outline-2">
532 <h2 id="sec-6"><span class="section-number-2">6</span> See also</h2>
533 <div class="outline-text-2" id="text-6">
535 Interesting or competing projects with good ideas:
539 <li><a href="#sec-3-2">CM-1 Connection Machine</a>
544 <li>database in the form of a knowledge graph that uses machine
545 reasoning to simplify data processing challenges for AI
546 applications. <a href="https://grakn.ai/">https://grakn.ai/</a>
551 <li><a href="http://wiki.squeak.org/squeak/2665">Magma</a>
553 <li>Multi-user object database for Squeak
558 <li><a href="http://esug.org/data/ESUG2015/3%20wednesday/1100-1130%20SQL%20Queries%20on%20Smalltalk%20Objects/SQL%20Queries%20in%20Smalltalk%20(James%20Foster).pdf">Gemstone/S</a>
560 <li>Completely distributed smalltalk based computing
566 <li><a href="http://www.uruk.org/emu/Taos.html">TAOS</a>
568 <li>Completely distributed operating system/virtual machine:
573 <li><a href="https://github.com/vygr/ChrysaLisp">ChrysaLisp</a>
575 <li>Assembler/C-Script/Lisp 64 bit, MIMD, multi CPU, multi threaded,
576 multi core, multi user Parallel OS. With GUI, Terminal, OO
577 Assembler, Class libraries, C-Script compiler, Lisp interpreter,
578 Debugger, and more…
585 </div><div class="col-md-3"><nav id="table-of-contents">
586 <div id="text-table-of-contents" class="bs-docs-sidebar">
588 <li><a href="#sec-1">1. General</a>
590 <li><a href="#sec-1-1">1.1. Source code</a></li>
593 <li><a href="#sec-2">2. Vision / goal</a></li>
594 <li><a href="#sec-3">3. Inspiration</a>
596 <li><a href="#sec-3-1">3.1. Brain</a></li>
597 <li><a href="#sec-3-2">3.2. CM-1 Connection Machine</a></li>
600 <li><a href="#sec-4">4. Ideas</a>
602 <li><a href="#sec-4-1">4.1. Geometrical computation</a></li>
603 <li><a href="#sec-4-2">4.2. Layered architecture</a></li>
606 <li><a href="#sec-5">5. Current status</a></li>
607 <li><a href="#sec-6">6. See also</a></li>
612 <footer id="postamble" class="">
613 <div><p class="author">Author: Svjatoslav Agejenko</p>
614 <p class="date">Created: 2019-09-20 Fri 09:53</p>
615 <p class="creator"><a href="http://www.gnu.org/software/emacs/">Emacs</a> 26.1 (<a href="http://orgmode.org">Org-mode</a> 9.1.9)</p>