|
Identifier |
000434592 |
Title |
Embedding key-value stores in object and database services |
Alternative Title |
Ενσωμάτωση συστημάτων κλειδιού-τιμής σε υπηρεσίες αντικειμένων και βάσεων δεδομένων |
Author
|
Καλαεντζής, Γεώργιος - Σάββας Σ.
|
Thesis advisor
|
Μπίλας, Άγγελος
|
Reviewer
|
Πρατικάκης, Πολύβιος
Μαγκούτης, Κώστας
|
Abstract |
Cloud storage is a storage model where data is stored through a cloud provider
that offers storage as a service with data durability, availability and security. All data and
configuration is held in the cloud and does not reside on the client site. As technologies
evolve, the size and amount of data that applications have to work with is growing
exponentially, data retrieval from the cloud impacts the performance of applications.
In particular, in the world of High-performance computing (HPC), data needs to be
retrieved as quickly as possible in order to perform complex calculations at high speeds.
Key-value stores are ideal for HPC environments, since they offer high performance and
faster application time.
In this thesis we aim to offer the benefits of cloud storage in an HPC environment.
We implement two systems, H3 and MongoKreon. H3 is a High speed, High Volume, and
High Availability system, with object storage semantics over high performance key-value
stores. MongoKreon is an integration layer between MongoDB and a state-of-the-art keyvalue store, Kreon. We describe the design and implementation of each system in detail
and present an evaluation analysis, compared to MinIO and Mongo-Rocks.
|
Language |
English |
Subject |
Cloud storage |
|
Object stores |
|
Αποθήκευση νέφους |
|
Συστήματα αποθήκευσης αντικειμένων |
|
Συστήματα αποθήκευσης κλειδιού τιμής |
Issue date |
2020-11-27 |
Collection
|
School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
|
|
Type of Work--Post-graduate theses
|
Views |
516 |