ABSTRAKSI: Akses terhadap cloud manager harus selalu dapat tersedia karena jika tidak maka akan menyebabkan user tidak dapat memonitor dan mengelola image, networking, privilige dan konfigurasi layanan IaaS cloud computing. Mengingat pentingnya akses terhadap cloud manager maka diperlukan suatu mekanisme agar cloud manager dapat selalu diakses meskipun server cloud manager tersebut mengalami down.
Solusi yang dapat diterapkan untuk mengantisipasi kegagalan layanan yang diakibatkan karena server cloud manager down yaitu dengan mekanisme fault tolerance. Dengan mekanisme fault tolerance diharapkan sistem dapat melanjutkan tugasnya dengan benar meskipun server cloud manager tersebut mengalami down. Oleh karena itu, pada tugas akhir ini diimplementasikan sebuah mekanisme fault tolerance pada cloud computing berbasis Eucalyptus menggunakan distibuted replicated block device (DRBD) dan Heartbeat.
Berdasarkan skenario server down, mekanisme fault tolerance menggunakan Heartbeat dan DRBD dapat diimplementasikan pada server cloud manager berbasis Eucalyptus. Kemudian, dari hasil percobaan dan analisis didapatkan nilai keepalive yang optimal adalah 0.2.Kata Kunci : fault tolerance, cloud manager, Eucalyptus, DRBD, Heartbeat.ABSTRACT: Access to cloud manager should always be available because otherwise it will lead to the user not be able to monitor and manage image, networking, and configuration privilige IaaS cloud computing services. Given the importance of access to the cloud manager, we need a mechanism for cloud manager could always be accessed though the manager of cloud server is down.
Solutions that can be applied to anticipate failures that caused by downed of cloud manager server is with fault tolerance mechanisms. Fault tolerance mechanisms is expected to continue the job of system properly despite the manager of cloud server is down. Therefore, in this research is implemented a fault tolerance in Eucalyptus-based cloud computing using distibuted replicated block device (DRBD) and Heartbeat.
Based on the scenario server down, fault tolerance mechanism using Heartbeat and DRBD can be implemented on the server-based cloud manager Eucalyptus. Then, from the results of experiments and analysis, writer conclude that the optimal value of keepalive is 0.2.Keyword: fault Tolerance, cloud manager, Eucalyptus, DRBD, Heartbeat.