Version 31
Created by tkirby on Jan 4, 2013 10:25 AM. Last modified by tkirby on Nov 29, 2013 10:14 AM.

    NEXT JBUG:

     

    Tuesday 10th December : JPUB - Venue the Hotspur.

     

    As it's Christmas we thought we'd have a social meetup. So don't turn up at INEX because we'll be in the Hotspur from 5:30 pm (or earlier), hope to see you there.

    For those of you not familiar with the Hotspur it is the pub opposite  the centre of the Haymarket bus station.

    see http://www.pubsnewcastle.co.uk/Hotspur.html

     

     

    Venue and Times

    The JBUG meetings are held in the Inex building. Start time is 6PM.  There will be someone to let you in between 5:45 PM  and 6:10 PM. If you need access out of those times can you let me know beforehand or you may not be able to gain access.

    Main talks usually finish around 8PM when Refreshments are provided.

     

    Details on how to find the venue can be found here.

     

    Upcoming JBUG Meetings

     

     

    Past JBUG Meetings

     

    Tuesday 12th November: Ryan Emerson: Improving Infinispan's Scalability: A new approach to group communication

     

    Signup for the event here: http://bit.ly/19fa2Th

     

    Infinispan is a distributed in memory key/value store that aims to be highly available and scalable.  Its ability to scale is tightly linked to the underlying order protocol, aka atomic broadcast, used for enforcing serializability.

     

    In this talk I will explore the scaling limits of Infinispan’s existing protocol and then propose an alternative approach to designing broadcast systems, that decouples  ordering and broadcast.

     

    This talk contrasts our approach with Infinispan’s, exploring the design features of our new approach and how they will improve Infinispan’s scalability.

     

     

    Tuesday 8th October - Jonathan Halliday: Travels with Cassandra.

     

    Leaving the comforts of SQL behind, we venture out into the wild, distributed, post-relational data frontier. No longer can we pick a ready to run query off the shelf. Out here if you want answers you have to roll up your sleeves and implement the algorithms, all the while keeping a sharp eye out for network partitions and fending off marauding node failures. On the up side, you can still get Pizza delivery and a decent cell phone signal.

     

     

    Tuesday Sept 10th - Derick James, CEO, Contemplate Ltd: Threadsafe

     

    Hangout on-air available here:  https://plus.google.com/b/101375244267574030408/events/c8h87g6rfkhaqilftcshfbl9ieo

     

    ThreadSafe is a static analysis product for finding concurrency bugs in multi-threaded Java programs.  Such bugs can be difficult to find using conventional testing methods.  They may occur randomly due to non-deterministic scheduling behaviour, and can be hard to replicate. ThreadSafe analyses can be from within the Eclipse IDE, and/or in conjunction with the Sonar quality dashboard.

     

    ThreadSafe is developed by Contemplate, a spin-out company from the University of Edinburgh.  Following its commercial release, it has been installed by an investment bank that participated in Contemplate's Early Adopter Program, and is under evaluation in others. The design intent is not only to identify the subtle issues which can arise in multi-threaded software development, but also to provide an easy to use user experience which allows the developer to quickly understand the analysis and to take remedial action, where required, with confidence and understanding.

     

    Contemplate will talk about its focus on concurrency, and demonstrate ThreadSafe's analysis of some Open Source examples.

     

     

    Tuesday August 13th - Alex Creasy: Transaction Monitoring and Visualisation

                                         Rui Vieira:     Efficient top-k query processing on distributed column family databases

     

     

    To make up for last months gap, we have two talks for you this month:

     

    Hangout on-air available here: https://plus.google.com/b/101375244267574030408/events/cl37lmktrfnpd0ospikv3sjjnsc

     

    Transaction Monitoring and Visualisation

     

    ACID Transactions are routinely used when applications require strong guarantees as to how atomic operations, involving multiple resources, will perform in the presence of failures, such as system crashes and network disruption. Transactional middleware such as JBoss, combined with application frameworks like JEE provide the developer with a simple means to add transactional semantics to their applications. Problems can still arise, however, for example poor transactional throughput may manifest when a high volume of transactions rollback, which can have a myriad of non obvious causes.

     

    This talk will explore the current methods of troubleshooting some common transactional issues using JBoss and introduce TxVis: a prototype transaction profiling and visualisation tool. We will discuss the challenges of its development and how it will aid the user in profiling the performance of transactions in their software and quickly isolate some commonly occurring problems.


     

    Efficient top-k query processing on distributed column family databases

     

    Ranking queries are one of the central topics in the field of Information Retrieval with considerable applications in fields such as analytics.

    One of the challenges is to provide solutions which can be adapted to distributed data sources, specifically NoSQL distributed column-oriented databases, and comply to "user" real-time constraints, especially when dealing with massive amounts of data.

    In this talk, we discuss the implementation and challenges for some of the most promising algorithms to address these challenges.

    We also analyse their scalability and substantial gains in terms of bandwidth and execution time, as indicated by the experimental results."


    Tuesday June 11th - Maciej Machulak, CEO, Cloud Identity Limited: Users in control of their data - User-Managed Access

     

    The amount of personal information stored on the Web is increasing. Such information is distributed and is often duplicated, incomplete or out-dated, and cannot be shared selectively by end-users. Personal Data Stores, where users input data once and use it many times with other parties, address the problem of managing distributed information. These systems, however, require authorization to allow secure data sharing. User-Managed Access (UMA), a standardisation effort incubated at Kantara Initiative and standardized at IETF, is a new approach to access control. It allows end-users to share Web data selectively. Users can easily determine what information will be revealed to which parties, or enforce how such data will be handled. In this talk, we discuss UMA and explain how it builds on OAuth 2.0, OpenID Connect and other Web standards. We will show an UMA Authorization Manager as well as applications built using open-source Python and UMA frameworks. Both frameworks can be used for flexible authorization in Web applications to allow end-users to share their data with other users and services on the Web. We will also show the Connect platform that simplifies federated authentication for such protocols as OAuth 1.0, OAuth 2.0, OpenID, OpenID Connect, and SAML, among many others.

     

    Maciej Machulak has worked in the field of Identity and Access Management for many years, both in industry and academia. He specializes in Security, Privacy and Trust on the Web, Cloud Security, Digital Identity and Software Engineering. Maciej serves as the Vice-Chair of the User-Managed Access Work Group at Kantara Initative and is one of the specification editors of the User-Managed Access (UMA) protocol for web-based access management.

    Prior to founding Cloud Identity, Maciej has worked for such companies as Red Hat and Google and contributed to the Apache Software Foundation. Maciej has also led efforts of the JISC/HEFCE funded SMART project that designed and implemented an UMA-based authorisation solution and deployed the system at Newcastle University. Maciej and his team has won the prestigious Kantara Initiative Identity Deployment of the Year (IDDY) 2011 Award for its efforts. His work has been presented at various international conferences, including JavaOne, Devoxx, MW4SOC, DIM, ICDS-SPCC, and IIW.

    Maciej holds an Advanced MSc degree in “System Design for Internet Applications” (SDIA) from Newcastle University and an MSc degree in “Secure Computer Networks” from Wroclaw University of Technology in Poland. Maciej is also finishing his PhD degree in the field of Web Security at Newcastle University in UK.

     

    Tuesday 14th May - Paul Robinson: Compensating Transactions: When ACID is too much

     

    ACID transactions are a useful tool for application developers and can provide very strong guarantees, even in the presence of failures. However, ACID transactions are not always appropriate for every situation.

     

     

    ACID transactions are achieved by holding locks on resources and require the ability to rollback changes. In some situations, the blocking nature of the protocol can be too limiting to performance, especially if the transaction is distributed. Also, some actions can’t simply be rolled back; for example, sending an email.

     

     

    A common strategy for applications that can’t use ACID, is to throw out transactions altogether. However, with this approach you are missing out on many of the benefits that transactions can provide. There are alternative transaction models that relax some of the ACID properties, while still retaining many of the strong guarantees essential for building robust enterprise applications. These should be considered before deciding not to use transactions at all.

     

    In this talk, I’ll present one such alternative to ACID transactions: compensation-based transactions. I’ll provide an overview of ACID and it’s limitations and describe some use-cases where ACID is not appropriate. I’ll then present our new API for compensation-based transactions and demonstrate how it can be used to address these problems. Finally, I’ll present our future plans to improve and expand our support for compensation-based transactions.

     

     

    Tuesday 9th April - Martyn Taylor: Utilizing multiple IaaS clouds using Deltacloud and Image Factory

    Talk Description:

     

    This session will consist of two talks given at FOSDEM 13 that address some of the issues with IaaS migration and vendor lock in.

     

    1st Talk

     

    Deltacloud:

     

    Deltacloud API prevents you from cloud vendor-lockin and cloud API changes. With deltacloud you can speak up to 18 different cloud providers using one single API. Deltacloud now officialy support the CIMI API as a new industry standard for cloud computing and also non-officialy Amazon EC2 query API.

    2nd Talk

     

    Cloud Agnostic Image Management:

     

    In a world where Cloud IaaS providers are popping up on every street corner, it becomes increasingly important to prevent vendor lockin.  Several efforts are already underway in projects like DeltaCloud to help abstract away the differences in API's and allow consumers to be cloud agnostic.

     

    However, the desire for cloud agnosticism becomes increasingly challenging when we start thinking about managing images.  Not only do we have API differences, but also many cloud vendors require differing formats and importation methods, some requiring uploading of pre-built images, others snapshots of running instances.

     

    This talk addresses some of the issues surrounding image management and looks at how these issues multiply in a federated cloud environment.  The talk will introduce a tool called Aeolus Image Factory; a project that offers image management abstraction for multitude of Cloud vendors.  Aeolus Image Factory allows users to define images at a high level and have them built, pushed and registered with all the major cloud vendors.

     

     

    Speaker bio:
     

    Martyn Taylor is a senior software engineer at Red Hat Ltd.  He is currently a member and chair of the Aeolus technical committee and is responsible for Aeolus ReST API and Cloud Image Management integration.

     

    websites:

     

    http://www.aeolusproject.org/

    http://imgfac.org/

     

     

    Tuesday 12th March - Professor Paul Watson, Dr Simon Woodman and Dr Hugo Hiden: E-Science Central

    Dr Simon Woodman and Dr Hugo Hiden are researchers in the School of Computing Science at Newcastle University.  Their areas of interest are scalable data analysis, workflows and mining of provenance data.  As a vehicle for their research and to support a wide range of scientists they built e-Science Central; a cloud based system for data analysis.  e-SC, deployed on JBoss middleware can run in various cloud environments and supports £20M of research across 2 faculties in Newcastle University.  It has been shown to be applicable to a different disciplines from analysis of physical activity data to computational fluid dynamics.

     

    In the talk they will present an overview of e-SC, give examples of it's use both in academia and industry, examine the use of JBoss middleware to support e-SC and discuss the future of e-SC and a collaborative project with RedHat and Arjuna.

    e-Science Central is a Cloud based Platform for Data Analysis.

     

    Tuesday 12th February - Two talks

    Trevor Kirby and Sanne Grinovero: Java and the Raspberry Pi

     

    Recently a large number of small devices have appeared on the market that are easy to program and are cheap and small enough to used anywhere. Most of these can drive hardware directly and this coupled with low footprint software opens up a whole range of possibilities. Coupled with these developments there has been a massive explosion in pre-built sensor boards and example programs. These changes mean there are an endless number of possibilities to explore using these boards.

     

    This talk will use the raspberry pi to demonstrate some of the ways you can use these embedded devices and give you an insight into some of the  technologies and possibilities.

     

    Jesper Pedersen: What is Java EE Connector Architecture? ... and why should you even care?

    Java EE Connector Architecture is one of the "hidden" technologies in the Java EE platform, but it is actually quite important.

     

    In this session we will take a look at what Java EE Connector Architecture is, and how to get started with using its powers.

     

    Agenda:

    • What is Java EE Connector Architecture ?
    • Hands-on
    • IronJacamar
    • JBoss Application Server 7

     

    Bio:

     

    Jesper Pedersen works as a Core Developer at Red Hat, where he is leading the IronJacamar project. IronJacamar provides the Java EE Connector Architecture implementation for JBoss Application Server.

     

     


    2187 Views Categories: Tags: