Rui Oliveira

All companies developing their business on the Web, not only giants like Google or Facebook but also small com- panies focused on niche markets, face scalability issues in data management. The case study of this paper is the content management systems for classified or commercial advertise-ments on the Web. The data involved has a very significant growth rate and a read-intensive access pattern with a reduced update rate. Typically, data is stored in traditional file systems hosted on dedicated servers or Storage Area Network devices due to the generalization and ease of use of file systems. However, this ease in implementation and usage has a disadvantage: the centralized nature of these systems leads to availability, elasticity and scalability problems. The scenario under study, undemanding in terms of the system's consistency and with a simple interaction model, is suitable to a distributed database, such as Cassandra, conceived precisely to dynamically handle large volumes of data. In this paper, we analyze the suitability of Cassandra as a substitute for file systems in content management systems. The evaluation, conducted using real data from a produc- tion system, shows that using Cassandra, one can easily get horizontal scalability of storage, redundancy across multiple independent nodes, and load distribution imposed by the periodic activities of safeguarding data, while ensuring a comparable performance to that of a file system.

Today's intensive demand for data such as live broadcast or news feeds requires effcient and robust dissemination systems. Traditionally, designs focus on extremes of the effciency/robustness spectrum by either using structures, such as trees for effciency or by using loosely-coupled epidemic protocols for robustness. We present Brisa, a hybrid approach combining the robustness of epidemics with the effciency of structured approaches. Brisa implicitly emerges embedded dissemination structures from an underlying epidemic substrate. The structures' links are chosen with local knowledge only, but still ensuring connectivity. Failures can be promptly compensated and repaired thanks to the epidemic substrate, and their impact on dissemination delays masked by the use of multiple independent structures. Besides presenting the protocol design, we conduct an extensive evaluation in real environments, analyzing the effectiveness of the structure creation mechanism and its robustness under dynamic conditions. Results confirm Brisa as an effcient and robust approach to data dissemination in large dynamic environments.

NoSQL databases were initially devised to support a few concrete extreme scale applications. Since the specificity and scale of the target systems justified the investment of manually crafting application code their limited query and indexing capabilities were not a major im- pediment. However, with a considerable number of mature alternatives now available there is an increasing willingness to use NoSQL databases in a wider and more diverse spectrum of applications and, to most of them, hand-crafted query code is not an enticing trade-off. In this paper we address this shortcoming of current NoSQL databases with an effective approach for executing SQL queries while preserving their scalability and schema flexibility. We show how a full-fledged SQL engine can be integrated atop of HBase leading to an ANSI SQL compli- ant database. Under a standard TPC-C workload our prototype scales linearly with the number of nodes in the system and outperforms a NoSQL TPC-C implementation optimized for HBase.

There is an increasing demand for efficient and robust systems able to cope with today's global needs for intensive data dissemination, e.g., media content or news feeds. Unfortunately, traditional approaches tend to focus on one end of the efficiency/robustness design spectrum, by either leveraging rigid structures such as trees to achieve efficient distribution, or using loosely-coupled epidemic protocols to obtain robustness. In this paper we present BRISA, a hybrid approach combining the robustness of epidemic-based dissemination with the effi- ciency of tree-based structured approaches. This is achieved by having dissemination structures such as trees implicitly emerge from an underlying epidemic substrate by a judicious selection of links. These links are chosen with local knowledge only and in such a way that the completeness of data dissemination is not compromised, i.e., the resulting structure covers all nodes. Failures are treated as an integral part of the system as the dissemination structures can be promptly compensated and repaired thanks to the underlying epidemic substrate. Besides presenting the protocol design, we conduct an extensive evaluation in a real environment, analyzing the effectiveness of the structure creation mechanism and its robustness under faults and churn. Results confirm BRISA as an efficient and robust approach to data dissemination in the large scale.

CumuloNimbo aims at solving the lack of scalability of transactional applications that represent a large fraction of existing applications. CumuloNimbo aims at conceiving, architecting and developing a transactional, coherent, elastic and ultra scalable Platform as a Service. Its goals are: Ultra scalable and dependable, able to scale from a few users to many millions of users while at the same time providing continuous availability; Support transparent migration of multi-tier applications (e.g. Java EE applications, relational DB applications, etc.) to the cloud with automatic scalability and elasticity. Avoid reprogramming of applications and non-transparent scalability techniques such as sharding. Support transactions for new data stores such as cloud data stores, graph databases, etc.The main challenges are: Update ultrascalability (million update transactions per second and as many read-only transactions as needed). Strong transactional consistency. Non-intrusive elasticity. Inexpensive high availability. Low latency. CumuloNimbo goes beyond the state of the art by scaling transparently transactional applications to very large rates without sharding, the current practice in Today?s cloud. In this paper we describe CumuloNimbo architecture and its performance.

The sheer volumes of data handled by today's Internet services demand uncompromising scalability from the persistence substrates. Such demands have been successfully addressed by highly decentralized key-value stores invariably governed by a distributed hash table. The availability of these structured overlays rests on the assumption of a moderately sta- ble environment. However, as scale grows with unprecedented numbers of nodes the occurrence of faults and churn becomes the norm rather than the exception, precluding the adoption of rigid control over the network's organization. In this position paper we outline the major ideas of a novel architecture designed to handle today's very large scale demand and its inherent dynamism. The approach rests on the well-known reliability and scalability properties of epidemic protocols to minimize the impact of churn. We identify several challenges that such an approach implies and speculate on possible solutions to ensure data availability and adequate access performance.

Massive-scale distributed computing is a challenge at our doorstep. The current exponential growth of data calls for massive-scale capabilities of storage and processing. This is being acknowledged by several major Internet players embracing the cloud computing model and offering first generation distributed tuple stores. Having all started from similar requirements, these systems ended up providing a similar service: A simple tuple store interface, that allows applications to insert, query, and remove individual elements. Further- more, while availability is commonly assumed to be sustained by the massive scale itself, data consistency and freshness is usually severely hindered. By doing so, these services focus on a specific narrow trade-off between consistency, availability, performance, scale, and migration cost, that is much less attractive to common business needs. In this paper we introduce DataDroplets, a novel tuple store that shifts the current trade-off towards the needs of common business users, pro- viding additional consistency guarantees and higher level data process- ing primitives smoothing the migration path for existing applications. We present a detailed comparison between DataDroplets and existing systems regarding their data model, architecture and trade-offs. Prelim- inary results of the system's performance under a realistic workload are also presented.

In this paper we review what it has been stated so far about transactional support on the cloud computing environment. Then, we propose to extend them with some ideas already stated in replicated databases, like the certification process, to solve certain problems about coordination in the commit phase of transactions in the cloud.