-
Notifications
You must be signed in to change notification settings - Fork 186
/
Copy pathindex.html
341 lines (306 loc) · 13.5 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
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
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
<!DOCTYPE html>
<html lang="en">
<head>
<title>ClickHouse: New Horizons</title>
<meta charset="utf-8">
<meta http-equiv="x-ua-compatible" content="ie=edge">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" href="shower/themes/yandex/styles/screen-16x9.css">
<style type="text/css">
code { display: block; white-space: pre; background-color: #EEE; }
p.cloud { text-align: center; line-height: 1.6; font-size: 12pt; }
p.cloud span { color: gray; padding: 0 20px 0 20px; white-space: nowrap; }
</style>
</head>
<body class="shower list">
<header class="caption">
<h1>ClickHouse: New Horizons</h1>
</header>
<section class="slide" id="cover" style="background: linear-gradient(to bottom right, rgba(255, 51, 51, 1), rgba(255, 169, 0, 1))">
<img src="https://presentations.clickhouse.com/original_website/logo.svg" style="width: 15%; position: absolute; right: 20px; bottom: 20px;" />
<h1 style="margin-top: 150px; line-height: 150%; color: white; font-size: 64pt">ClickHouse:<br/>New Horizons</h1>
</section>
<section class="slide">
<h2>The Origins</h2>
<p>2009 — first prototype of ClickHouse.</p>
<p>2012 — first production usages in Yandex for logs processing.</p>
<p>2014 — first production usages in Yandex for public service.</p>
<p>2014..2015 — rising adoption in Yandex.</p>
<p>2016 — ClickHouse release in <b>open-source</b>.</p>
</section>
<section class="slide" style="background: url('pictures/highload.png') no-repeat; background-size: 100%;">
</section>
<section class="slide">
<h2>The Origins</h2>
<p>2016..2017 — first usages outside of Yandex<br/>— for similar scenarios.</p>
<p>2018..2019 — distribution over the world<br/>— meetups in the US, Europe and <b>China</b>.</p>
</section>
<section class="slide" style="background: url('pictures/meetups.jpg') no-repeat; background-size: 100%;">
</section>
<section class="slide">
<h2>The Origins</h2>
<p>2020..2021:</p>
<p>Increased adoption of ClickHouse in large companies.<br/>
<span style="color: gray;">Due to increase of traffic and shortage of hardware.</span></p>
<p>ClickHouse appeared in many cloud services.<br/>
<span style="color: gray;">Alibaba, Tencent, Huawei, Bytedance, JD, Altinity, Yandex, Mail.ru, Sber...</span></p>
<p>Companies with services around ClickHouse.<br/>
<span style="color: gray;">Hydrolix, Tinybird, Gigapipe, LogTail...</span></p>
</section>
<section class="slide" style="background: url('pictures/discussion.jpg') no-repeat; background-size: 100%;">
<h2 style="color: white; text-shadow: -1px -1px 0 #000, 1px -1px 0 #000, -1px 1px 0 #000, 1px 1px 0 #000;">ClickHouse, Inc.</h2>
</section>
<section class="slide">
<h2>What Is ClickHouse, Inc?</h2>
<p><b>International, <span style="color: green;">independent</span> company.</b></p>
<p>Headquarters in Amsterdam.</p>
<p>Distributed team. We are hiring in Europe, Asia-Pacific and all over the world.</p>
</section>
<section class="slide">
<h2>ClickHouse, Inc</h2>
<p style="margin-top: -1em;"><b>Aaron Katz</b>, CEO - ex. Elastic, Salesforce.</p>
<p style="margin-top: -1em;"><b>Alexey Milovidov</b>, CTO - ex. Yandex.</p>
<p style="margin-top: -1em;"><b>Yuri Izrailevsky</b>, President of Engineering - ex. Google, Netflix.</p>
<div style="float: right; line-height: 0; margin-right: -50px;">
<img style="width: 100px;" src="pictures/people1.jpg"/><img style="width: 100px;" src="pictures/people2.jpg"/><img style="width: 100px;" src="pictures/people3.jpg"/><img style="width: 100px;" src="pictures/people4.jpg"/><img style="width: 100px;" src="pictures/people5.jpg"/>
<br/><img style="width: 100px;" src="pictures/people6.jpg"/><img style="width: 100px;" src="pictures/people7.jpg"/><img style="width: 100px;" src="pictures/people8.jpg"/><img style="width: 100px;" src="pictures/people9.jpg"/><img style="width: 100px;" src="pictures/people10.jpg"/>
<br/><img style="width: 100px;" src="pictures/people11.jpg"/><img style="width: 100px;" src="pictures/people12.jpg"/><img style="width: 100px;" src="pictures/people13.jpg"/><img style="width: 100px;" src="pictures/people14.jpg"/><img style="width: 100px;" src="pictures/people15.jpg"/>
<br/><b>...</b>
</div>
<img style="width: 50%; margin-left: -50px" src="pictures/founders.webp">
</section>
<!--
<section class="slide">
<h2>ClickHouse Mission</h2>
<p>1. Make ClickHouse the №1 analytical DBMS in the world.</p>
<p>2. ... everything else.</p>
</section>
-->
<section class="slide">
<h2>What's New In ClickHouse</h2>
<p>Season 2021/2022</p>
</section>
<section class="slide">
<h2>User Defined Functions and Types</h2>
<p style="margin-top: -1em;">1. SQL User Defined Functions (since 21.10)</p>
<code style="line-height: 1.5;">CREATE FUNCTION ch AS (x) -> (exp(x) + exp(-x)) / 2;
CREATE FUNCTION getRegionName AS (x)
-> dictGet('region', 'name', x);</code>
<p style="margin-top: 1em;">2. Scriptable User Defined Functions (since 21.11)</p>
<code style="font-size: 14pt; line-height: 1.1"><function>
<type>executable</type>
<name><b>myPlus</b></name>
<return_type>UInt64</return_type>
<argument><type>UInt64</type></argument>
<argument><type>UInt64</type></argument>
<format>TabSeparated</format>
<command><b>awk '{ print $1 + $2 }'</b></command>
<execute_direct>0</execute_direct>
</function>
</code>
</section>
<section class="slide">
<h2>Support For Semistructured Data</h2>
<p><b>JSON</b> data type:</p>
<code>CREATE TABLE games (data JSON) ENGINE = MergeTree;</code>
<p style="margin-top: 1em;">You can insert arbitrary nested JSONs.</p>
<p>Types are automatically inferred on INSERT and merge.</p>
<p>Data is stored in columnar format: columns and subcolumns.</p>
<p>Query nested data naturally.</p>
</section>
<section class="slide">
<h2>Support For Semistructured Data</h2>
<p style="margin-top: -1em;">Example: NBA games dataset</p>
<code>CREATE TABLE games (data String)
ENGINE = MergeTree ORDER BY tuple();
SELECT <span style="color: red;">JSONExtractString(data, 'teams', 1, 'name')</span>
FROM games;</code>
<p style="margin-top: 0.5em; color: red;">— 0.520 sec.</p>
<code>CREATE TABLE games (data JSON)
ENGINE = MergeTree;
SELECT <span style="color: green;">data.teams.name[1]</span> FROM games;</code>
<p style="margin-top: 0.5em; color: green;">— 0.015 sec.</p>
</section>
<section class="slide">
<h2>Support For Semistructured Data</h2>
<code style="font-size: 12pt; margin-top: -2em;">DESCRIBE TABLE games
SETTINGS describe_extend_object_types = 1
name: data
type: Tuple( <-- inferred type
`_id.$oid` String,
`date.$date` String,
`teams.abbreviation` Array(String),
`teams.city` Array(String),
`teams.home` Array(UInt8),
`teams.name` Array(String),
`teams.players.ast` Array(Array(Int8)),
`teams.players.blk` Array(Array(Int8)),
`teams.players.drb` Array(Array(Int8)),
`teams.players.fg` Array(Array(Int8)),
`teams.players.fg3` Array(Array(Int8)),
`teams.players.fg3_pct` Array(Array(String)),
`teams.players.fg3a` Array(Array(Int8)),
`teams.players.fg_pct` Array(Array(String)),
`teams.players.fga` Array(Array(Int8)),
`teams.players.ft` Array(Array(Int8)),
`teams.players.ft_pct` Array(Array(String)),
`teams.players.fta` Array(Array(Int8)),
`teams.players.mp` Array(Array(String)),
`teams.players.orb` Array(Array(Int8)),
`teams.players.pf` Array(Array(Int8)),
`teams.players.player` Array(Array(String)),
`teams.players.plus_minus` Array(Array(String)),
`teams.players.pts` Array(Array(Int8)),
`teams.players.stl` Array(Array(Int8)),
`teams.players.tov` Array(Array(Int8)),
`teams.players.trb` Array(Array(Int8)),
`teams.results.ast` Array(Int8),
`teams.results.blk` Array(Int8),
`teams.results.drb` Array(Int8),
`teams.results.fg` Array(Int8),
`teams.results.fg3` Array(Int8),
`teams.results.fg3_pct` Array(String),
`teams.results.fg3a` Array(Int8),
`teams.results.fg_pct` Array(String),
`teams.results.fga` Array(Int16),
`teams.results.ft` Array(Int8),
`teams.results.ft_pct` Array(String),
`teams.results.fta` Array(Int8),
`teams.results.mp` Array(Int16),
`teams.results.orb` Array(Int8),
`teams.results.pf` Array(Int8),
`teams.results.plus_minus` Array(String),
`teams.results.pts` Array(Int16),
`teams.results.stl` Array(Int8),
`teams.results.tov` Array(Int8),
`teams.results.trb` Array(Int8),
`teams.score` Array(Int16),
`teams.won` Array(Int8))</code>
</section>
<section class="slide">
<h2>Support For Semistructured Data</h2>
<p>Flexible schema.</p>
<p>You can have columns with strict and flexible schema in one table.</p>
<p>Queries work <span style="color: green;">as fast as</span> with predefined types!</p>
<p style="color: gray; margin-top: 5em;">Planned for Q1 2022.</p>
</section>
<section class="slide">
<h2>Schema Inference From Formats</h2>
<p style="margin-top: -1em;">Before:</p>
<code>clickhouse-local --input-format <b>Parquet</b> \
--query 'SELECT * FROM table' \
<span style="color: red;">--structure</span> '
id UInt32, deleted UInt8,
type String, by String, time DateTime,
text String, dead UInt8, parent UInt32,
poll UInt32, kids Array(UInt32),
url String, score Int32, title String,
parts Array(UInt32), descendants Int32'
</code>
<p style="margin-top: 1em;">After:</p>
<code>clickhouse-local --input-format <b>Parquet</b> \
--query 'SELECT * FROM table'
</code>
</section>
<section class="slide">
<h2>Schema Inference From Formats</h2>
<p>1. No need to specify schema if data already contains it:<br/>
Native, Arrow, Parquet, ORC, Avro.</p>
<p>2. Allow to infer schema from text formats:<br/>
TSV, CSV, JSONEachRow.</p>
<p>It works for:</p>
<p>— clickhouse-local;
<br/>— url;
<br/>— file;
<br/>— s3.</p>
<p style="color: gray; margin-top: 1em;">Planned for Q1 2022.</p>
</section>
<section class="slide">
<h2>Text Classification and NLP</h2>
<p style="line-height: 1.5">1. Tokenization, stemming, lemmatization, synonims (21.9..21.11).
<br/>— SELECT arrayMap(word -> lemmatize(word), tokens(text)).</p>
<p>2. Charset and language detection (Q4 2021).</p>
<p style="line-height: 1.5">3. Semi-duplicate text search (21.1).
<br/>— min-hash and sim-hash algorithms.</p>
</section>
<section class="slide">
<h2>ZooKeeper Will Go Away</h2>
<p>1. clickhouse-keeper — 100% compatible protocol with ZooKeeper.</p>
<p>— compressed logs and snapshots;
<br/>— no issues with zxid overflow;
<br/>— no issues with large packets;
<br/>— better memory usage;
<br/>— no issues with GC and Java heap;</p>
<p>2. Can run <b>embedded</b> into clickhouse-server.</p>
<p>— no need for a separate process;</p>
<p style="color: gray">Preproduction stage, planned for Q4 2021.</p>
</section>
<section class="slide">
<h2>Asynchronous INSERTs</h2>
<p>Allow to do many small INSERTS.</p>
<p>From many concurrent connections.</p>
<p style="color: green">Without Kafka and Buffer tables!</p>
<p>Many small INSERTs are combined together<br/>
and batched in memory.</p>
<p>Safe by default: every client get response when data is inserted.</p>
<p style="color: gray;">Available in 21.11-testing</p>
</section>
<section class="slide">
<h2>Support For Transactions</h2>
<p>ClickHouse is not a transactional DBMS, isn't it?</p>
<p>But we need transactions to:</p>
<p>— do atomic INSERTs to multiple partitions;</p>
<p>— do atomic INSERTs to multiple tables and materialzied views;</p>
<p>— make multiple SELECT queries from one snapshot.</p>
<p style="color: gray;">Planned for Q2 2022.</p>
</section>
<section class="slide">
<h2>Key-Value Data Marts</h2>
<p>... And Incremental Aggregation In Memory</p>
<p>Run GROUP BY query continuously.</p>
<p>Aggregated data is accumulated in memory.</p>
<p>It can be queried as a table.</p>
<p>The server can serve key-value requests with Redis protocol.</p>
<p><span style="color: green;">Example applications:</span>
<br/>— realtime antifraud;
<br/>— user profiles and personalization.</p>
<p style="color: gray;">Planned for summer 2022.</p>
</section>
<section class="slide">
<h2>Decoupled Storage</h2>
<p>ClickHouse over S3:</p>
<p>— highly concurrent asynchronous reads (21.11);</p>
<p>— parallel processing on multiple compute nodes (Q4);</p>
<p>— caching in RAM or local disk with node affinity (Q4);</p>
<p>— easy coordination of inserts and merges<br/>  without ReplicatedMergeTree (Q1 2022);</p>
<p style="color: green;">Lower cost and higher throughput than EBS volumes.</p>
<p>Bonus: "web" disk for static hosted datasets (21.10).</p>
</section>
<section class="slide">
<h2>What Else?</h2>
<p>Streaming queries:<br/>
— GROUP BY and ORDER BY on time frames;<br/>
— distributed streams;<br/>
— persistent streams;<br/>
— subscription to tables and external data streams;
</p>
<p>Advanced time series analysis.</p>
<p>Tamper-proof data storage.</p>
<p>Embedded ClickHouse.</p>
<p>ETL and batch jobs.</p>
</section>
<section class="slide">
<h2>Thank You!</h2>
<p>Please:</p>
<p>— Use ClickHouse.</p>
<p>— 💖 ClickHouse.</p>
<p>— Contribute to ClickHouse!</p>
</section>
<div class="progress"></div>
<script src="shower/shower.min.js"></script>
<!--Video plugin-->
<link rel="stylesheet" href="shower/shower-video.css">
<script src="shower/shower-video.js"></script>
<!--/Video plugin-->
</body>
</html>