Panagiotis Garefalakis

In this paper 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, which records data mutations to a Log-structured Merge Tree (LSM-Tree). ACaZoo scales by horizontally partitioning the key space via… Show more

In this paper 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, which records data mutations to a Log-structured Merge Tree (LSM-Tree). ACaZoo scales by horizontally partitioning the key space via consistent primary-key hashing on available 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 40% in write-intensive workloads. 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. Show less

Cloud computing is becoming a popular platform to deliver service-based applications (SBAs) based on service-oriented architecture (SOA) principles. Monitoring the performance and functionality of SBAs deployed on multiple Cloud providers (in what is also known as Multi-Cloud setups) and adapting them to variations/events produced by several layers (infrastructure, platform, application, service, etc.) in a coordinated manner are challenges for the research community. This paper proposes a… Show more

Cloud computing is becoming a popular platform to deliver service-based applications (SBAs) based on service-oriented architecture (SOA) principles. Monitoring the performance and functionality of SBAs deployed on multiple Cloud providers (in what is also known as Multi-Cloud setups) and adapting them to variations/events produced by several layers (infrastructure, platform, application, service, etc.) in a coordinated manner are challenges for the research community. This paper proposes a monitoring framework for Multi-Cloud SBAs with two main objectives: (a) perform cross-layer (Cloud and SOA) monitoring enabling concerted adaptation actions; (b) address new challenges raised in Multi-Cloud SBA deployment. The proposed framework is empirically evaluated on a real-world Multi-Cloud setup. Show less

Increasingly larger data centers constructed to serve Internet-scale enterprise and Cloud computing workloads are creating significant challenges for the management systems designed to operate them. In this paper we present the design and implementation of a management system that uses server-attached wireless sensors to create an auto-configuring wireless-only monitoring network that can be used to address a number of challenges in the area of data center operations management. We exhibit the… Show more

Increasingly larger data centers constructed to serve Internet-scale enterprise and Cloud computing workloads are creating significant challenges for the management systems designed to operate them. In this paper we present the design and implementation of a management system that uses server-attached wireless sensors to create an auto-configuring wireless-only monitoring network that can be used to address a number of challenges in the area of data center operations management. We exhibit the use of this wireless network in tracking the physical location of servers and in troubleshooting connectivity problems in the wired data center network infrastructure. Our system is implemented as an extension to the Nagios distributed monitoring and management system and demonstrated to provide continuous awareness of the physical location of the server infrastructure as well as insight into wired network partitions under switch failures. Show less

Distributed highly-available key-value stores have emerged as important building blocks for applications handling large amounts of data. The Apache Cassandra system is one such popular store combining a key distribution mechanism based on consistent hashing with eventually-consistent data replication and membership mechanisms. Cassandra fits well applications that share its semantics but is a poor choice for traditional applications that require strong data consistency. In this work we… Show more

Distributed highly-available key-value stores have emerged as important building blocks for applications handling large amounts of data. The Apache Cassandra system is one such popular store combining a key distribution mechanism based on consistent hashing with eventually-consistent data replication and membership mechanisms. Cassandra fits well applications that share its semantics but is a poor choice for traditional applications that require strong data consistency. In this work we strengthen the consistency of Cassandra through the use of appropriate components: the Oracle Berkeley DB Java Edition High Availability storage engine for data replication and a replicated directory for maintaining membership information. The first component ensures that data replicas remain consistent despite failures. The second component simplifies Cassandra’s membership, improving its consistency and availability. We argue that the resulting system fits a wider range of applications, and is more robust and easier to reason about. Show less