SysFera-DS was first developed in the GRAAL INRIA project team from the Laboratoire de l’Informatique du Parallelisme of the Ecole Normale Superieure de Lyon during 9 years now. This software allows you to get an IT infrastructure that is coherent, efficient and transparent for the user, from both hardware and software points of view.

SysFera-DS is the commercial version the DIET (Distributed Interactive Engineering Toolbox) software.

SysFera-DS is focused on the development of scalable middleware with initial efforts focused on the distribution of the scheduling across multiple agents. SysFera-DS consists of a set of elements that can be used together to build applications using the GridRPC paradigm in a ASP (Application Service Provider) or SaaS (Sofware as a Service) way which simplifies the usage of a set of resources (computing, storage, network, etc.) This middleware is able to find an appropriate server according to the information given in the client's request (e.g. problem to be solved, size of the data involved), the performance of the target platform (e.g. server load, available memory, communication performance) and the local availability of data stored during previous computations. The scheduler is distributed using several collaborating hierarchies connected either statically or dynamically (in a peer-to-peer fashion.) Data management is provided to allow persistent data to stay within the system for future re-use. This feature avoids unnecessary communication when dependencies exist between different requests.
SysFera-DS has the following functionnalities:

• Adaptive scheduling framework. Applications targeted by the SysFera-DS platform are able to exert a degree of control over the scheduling subsystem through plug-in schedulers. As the applications that are to be deployed on the grid vary greatly in terms of performance demands, the SysFera-DS plug-in scheduler facility allows the application designer to express application needs and features in order that they are taken into account when application tasks are scheduled.

  Usage: this functionality allows, for example, to select a computing resource fulfilling the application's memory requirements.

  Strength: the adaptive scheduling framework is a unique innovation, which does not exist within equivalent middleware. It allows a very strong dynamic adaptation to both applications and resources.

• Management of data/work/controlflows. A large number of applications can be represented by graphs of tasks connected with data dependencies, and with (or without) control instructions (such as loops, if statements, etc.) SysFera-DS is able, based on those representations to manage multiple control/data/workflows at the same time by executing the possibly parallel tasks, the data transfers and selecting the best suitable resources available at the execution time.

  Usage: in the case of medical imaging, the images' processing and analysis can be quite complex. This process is often represented as a workflow graph which need to be applied on all the available images.

  Strength: a unique characteristic of SysFera-DS, is that it provides load-balancing between several competing workflows. Only our solution provides this feature.

• Smart data management. Several applications have to deal with large amout of data that can be replicated to limit the transfer time between the client and the server. Data may need to be duplicated, replicated, moved or stored in special storage elements. In order to fulfill these requirements, SysFera-DS provides DAGDA, a efficient and powerful module allowing data explicit or implicit replications, and advanced data management.

  Usage: in genomic, very large genes databases need to be frequently accessed, thus generating large amounts of data to be transfered. A way to overcome this problem, and reduce the data movements, is to replicate at needs these databases on the site they are used (the computing servers.)

  Strength: The data manager allows implicit and explicit data replication. It also integrates intelligent mechanisms for storage space management, and transfer sources selection.

• Batch Schedulers interfaces. Parallel resources are generally accessible through reservation systems, also called batch systems. To execute their jobs, clients have to submit jobs via special mnemonics and chosen options (most of the time both are very different from one batch system to another) a script that contains the command line that launches the job. SysFera-DS is able to abstract these differences through the use of a transparent and consistent API and allows the user to address resources managed by different batch schedulers (LSF, PBS, LoadLeveler, SGE, OAR, etc.) at the same time without having to know how to deal with the underlying system.

  Usage: when the resources are heterogeneous and managed by several jobs management systems, using several computing resources can be tedious for the user. Our product provides an abstraction layer, which provides transparent access to resources, whatever the system managing them.

  Strength: this feature allows, among other things, to efficiently manage a set of dynamic resources, as would be the case with IaaS during an activity peak.

• Security/Authentication/Authorization. SysFera-DS provides different levels of security and authentication for the client side (simple/advanced authentication with users, groups and certificates), the server side (access rights definitions for users/groups per services), and for the data managed by the middleware (access rights, and localization of data on available resources.) Moreover, communications can be crypted inside the middleware, and unauthorized access can be prevented.

  Usage: this feature provides to our clients an access to computing resources which usage can easily be charged to a third party, but it also allows to trace the resources' usage, and thus prevent prohibited usage.

  Strength: we provide a Single Sign On (SSO) solution.