Datenbanken — die Qual der Wahl

In kaum einem anderen Bereich gab es in den letzten Jahren eine derartig rege Neu- und Weiterentwicklung wie bei Datenbanken. Hatte man vor einigen Jahren oft noch "seine" Datenbank, die bei jedem Projekt zum Einsatz kam, ist man heute mit einer Vielzahl an Ansätzen und Implementierungen konfrontiert.

Wir starten mit einem kurzen Blick auf den theoretischen Hintergrund von verteilten Systemen und Datenbanken im Speziellen. Anhand dessen wird dann die Unterteilung in traditionelle, relationale Datenbanken wie PostgreSQL und MySQL auf der einen, beziehungsweise den neueren NoSQL Systemen wie MongoDB, Redis, Cassandra oder ElasticSearch auf der anderen Seite klar.
Danach diskutieren wir mögliche Einsatzszenarien sowie Vor- und Nachteile einiger Datenbanken:
* Wieso SQL (wieder) in Mode ist
* Warum MongoDB's Dokumenten-Struktur so gut zur objektorientierten Programmierung passt
* Wie man mit Redis effizient Zugriffe erfassen kann
* Wieso Cassandra so skalierbar und ausfallssicher ist
* Wie Volltextsuche mit ElasticSearch funktioniert

Die richtige Wahl der Datenbank(en) wurde durch die Vielzahl an Möglichkeiten sicherlich nicht einfacher, aber dafür um so interessanter!

Transcript

1. Datenbanken die Qual der Wahl Philipp Krenn̴̴̴̴̴̴̴@xeraa
5. Infrastructure | Developer Advocate
6. ViennaDB Papers We Love Vienna
9. "relational model and normal form" ↓ "data independence"
10. Relational Algebra Book "Seven Databases in Seven Weeks"
12. Declarative Implementation detail of the RDBMS (Query Optimizer)
13. User Interaction Aggregated reports Concurrency, integrity, consistency, type-safety
14. Codd's 12 rules
16. From a Wide Variety of Databases to A Dark Age
18. http://blog.8thlight.com/uncle-bob/ 2012/05/15/NODB.html
19. Slow & inflexible
23. https://aphyr.com/posts/288-the- network-is-reliable
24. Beating the CAP Theorem Checklist Your ( ) tweet ( ) blog post ( ) marketing material ( ) online comment advocates a way to beat the CAP theorem. Your idea will not work. http://ferd.ca/beating-the-cap-theorem-checklist.html
25. Schema Flexibility http://pragprog.com/magazines/ 2012-05/beyond-the-bit-bucket
30. NoSQL Not Only SQL
33. Big Data Very broad, including NoSQL Offline data / data warehouse, Hadoop & Spark
35. http://db-engines.com/en/ranking
37. CouchDB, Couchbase, Elasticsearch,...
38. Ted Neward: ORM is "The Vietnam of Computer Science" http://blogs.tedneward.com/post/the-vietnam-of-computer- science/
39. @OneToMany(mappedBy = "destCustomerId") @ManyToMany @Fetch(FetchMode.SUBSELECT) @JoinTable(name = "customer_dealer_map", joinColumns = { @JoinColumn(name = "customer_id", referencedColumnName = "id")}, inverseJoinColumns = { @JoinColumn(name = "dealer_id", referencedColumnName = "id")}) private Collection<Client> dealers;
41. public abstract class EmployeeEntity { protected String name; } public class ManagerEntity extends EmployeeEntity { protected Boolean approveFunds; } public class WorkerEntity extends EmployeeEntity { protected Integer yearsExperience; }
42. RDBMS 1. Union table with (many) NULL values
43. RDBMS 2. Concrete instances without common queries
44. RDBMS 3. Base table JOINed with concrete instances
45. @Entity(value = "employee", noClassnameStored = false) public abstract class EmployeeEntity { @Id protected ObjectId id; protected String name; } public class ManagerEntity extends EmployeeEntity { protected Boolean approveFunds; } public class WorkerEntity extends EmployeeEntity { protected Integer yearsExperience; }
46. { "_id": ObjectId("5461c8bf9e2acf32ed50c079"), "className": "net.xeraa.morphia_demo.entities.ManagerEntity", "name": "Peter", "approveFunds": true } { "_id": ObjectId("524d9fe7e4b0f8bd3031f84e"), "className": "net.xeraa.morphia_demo.entities.WorkerEntity", "name": "Philipp", "yearsExperience": 10 }
49. Key-Value Store and more REmote DIctionary Server
50. Statistics (Login) Bitset HyperLogLog
51. Bitset First user bit 1, second user bit 2,... A million users ~123KB
52. HyperLogLog Unique elements in 12KB regardless of their number Error rate ~0,81% http://redis.io/commands/pfcount
53. Comparison HyperLogLog: constant size Bitset: exact & aggregates
54. Overall Simple Omni-present "glue"
56. Scalable & highly available Consistent Hashing
57. https://highlyscalable.wordpress.com/2012/09/18/distributed-algorithms-in-nosql-databases/
62. * Without read and write QUORUM
63. Eventually consistent was too hard, so we went with immediately inaccurate. — https://twitter.com/mysqlborat/status/621785755107524608
67. Indices http://use-the-index-luke.com
68. SQL Databases not only RDBMS
69. Cassandra Query Language (CQL) Google Query Language (GQL) Couchbase N1QL RethinkDB Query Language (ReQL) ArangoDB Query Language (AQL)
70. In the main, AQL is a declarative language. — https://www.arangodb.com/why-arangodb/sql-aql-comparison/
71. AQL FOR doc IN users FILTER doc._key == "philipp" RETURN doc
72. RDBMS features & maturity + NoSQL influence
74. Full-Text Search
76. Example These are <em>not</em> the droids you are looking for.
77. html_strip Char Filter These are not the droids you are looking for.
78. standard Tokenizer These̴are̴not̴the̴droids̴you̴are̴looking̴for
79. lowercase Token Filter these̴are̴not̴the̴droids̴you̴are̴looking̴for
80. stop Token Filter droids̴you̴looking
81. snowball Token Filter droid̴you̴look
82. Document Score Term Frequency (TF) / Inverse Document Frequency (IDF) Field-length norm https://www.elastic.co/guide/en/elasticsearch/guide/current/scoring-theory.html
83. Is it a database?
84. Conclusion
86. SQL: What data do I have? NoSQL: What question do I have?
87. But is it fast?
90. Thanks! Questions? Stickers @xeraa
91. Credits Schnitzel https://flic.kr/p/9m27wm Architecture https://flic.kr/p/6dwCAe Conchita https://flic.kr/p/nBqSHT No, SQL https://twitter.com/edd/status/ 400190499585544192