|
|
| Identifier |
000388403 |
| Title |
A distributed key-value store based on replicated LSM-trees |
| Alternative Title |
Ένα κατανεμημένο σύστημα αποθήκευσης δεδομένων βασισμένο σε ομοιοτυπία LSM-δέντρων |
|
Author
|
Γαρεφαλάκης, Παναγιώτης Ι.
|
|
Thesis advisor
|
Μπίλας, Άγγελος
|
|
Reviewer
|
Μαγκούτης, Κωνσταντίνος
Πλεξουσάκης, Δημήτριος
|
| Abstract |
Distributed highly-available key-value stores have emerged as important build¬ing blocks for data-intensive applications. Eventually-consistent versions of such stores have become popular due to their high availability ("always writeable") features; they are however unsuitable for many applications that require strong consistency. In this thesis we describe the design and implementation of ACaZoo, a key-value store that combines strong consistency with high performance and high availability. ACaZoo supports the popular column-oriented data model of Apache Cassandra and HBase. It implements strongly-consistent data replication using primary-backup atomic broadcast of a write-ahead log, recording data muta¬tions to a Log-structured Merge Tree (LSM-Tree). ACaZoo scales by horizontally partitioning the key space via consistent primary-key hashing over replica groups (RGs).
LSM-Tree compactions can hamper performance, especially when they take place at RG primaries. ACaZoo addresses this problem by changing RG leadership prior to heavy compactions, a method that can improve throughput by up to 60% in write-intensive workloads. To further improve response time ACaZoo uses client-side routing of requests, which is known to complicate the propagation of configuration changes to a large and dynamic client population. We address this problem by proposing an optimized version of ZooKeeper that can load-balance issuing of change notifications across its servers. We evaluate ACaZoo using the Yahoo Cloud Serving Benchmark (YCSB) and compare it to Oracle's NoSQL Database and to Cassandra providing serial consistency via an extension of the Paxos algorithm. We further evaluate application performance using CassMail, a scalable e-mail service, over both ACaZoo and Cassandra.
|
| Language |
English, Greek |
| Subject |
Distributed systems |
|
NoSQL databases |
|
Ομοιοτυπία δεδομένων |
| Issue date |
2014-11-21 |
| Collection
|
School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
|
|
|
Type of Work--Post-graduate theses
|
| Views |
403 |
A distributed key-value store based on replicated LSM-trees
