Presentation is loading. Please wait.

Presentation is loading. Please wait.

SimCity Building Blocks at the DICE team

Similar presentations


Presentation on theme: "SimCity Building Blocks at the DICE team"— Presentation transcript:

1 SimCity Building Blocks at the DICE team
Marian Bubak, Bartosz Baliś, Marek Kasztelnik, Maciej Malawski, Piotr Nowakowski {bubak,balis, {m.kasztelnik, Department of Computer Science and ACC Cyfronet AGH Krakow, PL dice.cyfronet.pl

2 Example application: Flood threat assessment
Scenario: levee-protected area endangered by flood due to high water levels The user selects an area for flood threat assessment

3 Flood threat assessment: our solution

4 Data sets GIS data (levees, sensor locations) Sensor data
Simulated sensor data Computation results Lots of metadata

5 (Virtual) Experiments
Support for conducting virtual experiments: Flooding a reservoir protected by an artificial levee Observation of a levee during heavy rain Simulation of a flooding scenario Experiment lifecycle Creating a new experiment and defining its context Collecting information during the experiment Concluding the experiment Reusing experiment results in future experiments

6 HyperFlow: programming and execution of workflow-based scientific applications
Innovative programming approach and enactment engine for scientific workflows Combines declarative workflow description with low-level programming in JavaScript / node.js for implementing workflow activities Simple and concise syntax + mainstream scripting language & runtime platform = increased programming productivity Based on a formal model of computation (Process Networks) Supports a rich set of complex workflow patterns <workflow.json> { "processes": [ {   "name": "Sqr",     "function": "sqr",     "type": "dataflow",     "parlevel": 0, // level of parallelism (unlimited)     "ordering": true, // ordering of results     "ins": [ "number" ],     "outs": [ "square" ]   }, {     "name": "Sum",     "function": "sum",     "ins": [ "square:3" ],     "outs": [ "sum" ]   } ], "signals": [ {     "name": "number",     "data": [ 1, 2, 3, 4, 5, 6 ]     "name": "square"     "name": "sum"   "ins": [ "number" ],   "outs": [ "sum" ] } <functions.js> function sqr(ins, outs, config, cb) {     var n = Number(ins.number.data[0]);     outs.square.data = [n * n];     cb(null, outs); } function sum(ins, outs, config, cb) {     var sum=0.0;     ins.square.data.forEach(function (n) { sum += n; });     outs[0].data = [ sum ]; B. Baliś, Increasing Scientific Workflow Programming Productivity with HyperFlow. In Proceedings of the 9th Workshop on Workflows in Support of Large-Scale Science (In Print).

7 SimCity Requirements for “CIS”
Start a simulation based on user input. Let an automated component start workflows within CIS with new parameter sets, and receive results asynchronously. Workflow generator is a kind of an automated components generating and starting workflows. Results are received asynchronously by the user interface The user specifies “the what” (in this case, which are should be evaluated), the system is automatically generated (“the how”) All functions should be accessible from a user-friendly interface that is not concerned with how something is computed but with what is computed.

8 SimCity Requirements for “CIS”
Start an external component (for parameter exploration) via an external (custom) user interface. Ideally, parameter exploration is exposed as a web service, so different algorithms can be started from a custom web interface. The parameter exploration algorithm should be modifiable or at least selectable by the user.

9 SimCity Requirements for “CIS”
Stage data to and from clusters or cloud infrastructure (virtual machines). For each job, the input and output files should end up at the correct places. Construct input files based on parameter sets. TRANSIMS, for example, works with control files to determine the parameters, but also with data files with parameters. Both would have to be editable from CIS.

10 SimCity Requirements for “CIS”
Automatically schedule workflows on cluster or cloud infrastructure. Preferably, the user should be able to select their own cluster. We have the Optimizer/Scheduler for this.

11 SimCity Requirements for “CIS”
Provide feedback to a component based on user input. Parameter exploration may be guided by the user. In this case, ideally, a web service is provided for real-time interaction with the parameter exploration component. CIS gives feedback to a component based on sensor data, which may then start new workflows. Towards manual steering: Workflow processes run continuously Workflow processes expose a REST API which could be invoked from the User Interface Message queue could be used for “steering” the component. Towards automated steering: We could add another component to perform the automated decision making We’re planning something similar in ISMOP, e.g. automatic increasing of sensor data collection frequency based on current values. New workflows can certainly be automatically generated and started, without the user intervention.

12 Hybrid cloud as a means of provisioning computing power for virtual experiments – the Atmosphere framework Cloud Management Portlets GUI host (provisions end-user features and access options) Provide GUI elements which enable service developers and end users to interact with the Atmosphere platform and create/deploy services on the available cloud resources Worker Node Worker Node Worker Node Worker Node 96 CPU cores 184 GB RAM 4 TB storage private IP space Head Node Worker Node Worker Node Worker Node Worker Node Image store OpenStack cloud site at ACC CYFRONET AGH Atmosphere Core Services Host user accounts Atmosphere Registry (AIR) available cloud sites services and templates Atmosphere Core Secure RESTful API (Cloud Facade) Authentication and authorization logic Communication with underlying computational clouds Launching and monitoring service instances Creating new service templates Billing and accounting Logging and administrative services Worker node w/large resource pool („fat node”) 128 CPU cores 256 GB RAM 4 TB storage private IP space Head Node Worker node w/large resource pool („fat node”) Image store VPH-Share cloud site at UNIVIE Worker Node Massive (functionally limitless) hardware resource pool public IP space API host Worker Node Image store Amazon Elastic Compute Cloud (EC2) – European availability zone

13 Atmosphere platform interfaces
End user A full range of user-friendly GUIs is provided to enable service creation, instantiation and access. A comprehensive online user guide is also available. The GUIs work by invoking a secure RESTful API which is exposed by the Atmosphere host. We refer to this API as the Cloud Facade. Atmosphere Registry (AIR) Atmosphere Ruby on Rails controller layer (core Atmosphere logic) Cloud sites Application -- or -- Workflow environment Any operation which can be performed using the GUI may also be invoked programmatically by tools acting on behalf of the platform user – this includes standalone applications and workflow management environments (which VPH-Share also provides). All operations on cloud hardware are abstracted by the Atmosphere platform which exposes a unified RESTful API (with a suitable set of developer’s documentation available). For end users, the API is concealed by a layer of platform GUIs embedded in the VPH-Share portal and providing a user-friendly work environment - for domain scientists and service developers alike. The API can also be directly invoked by external services as long as they possess the required security credentials (Atmosphere relies on the well-known OpenID authentication standard).

14 Shared and scalable services – smart utilization of hardware resources
Scientist Developer Atmosphere Cloud Platform Cloud Service Published Published services become visible to non-developers and can be instantiated using the Generic Invoker. Developers are free to spawn „snapshot” images of their cloud services (e.g. for backup purposes) without exposing them to external users. A Shared service is backended by a single virtual machine which „mimics” multiple instances from the users’ point of view. Shared services greatly conserve hardware resources and can be instantiated quickly. Atmosphere Cloud Service Shared Cloud WN Shared VM Scientist When a Scalable service is overloaded with requests Atmosphere can spawn additional instances in the cloud to handle the additional load. The process is transparent from the user’s perspective. Atmosphere Cloud Service Scalable Cloud WN Separate VM Scientist

15 More information about the hybrid computational cloud platform
A more detailed introduction to the Atmosphere cloud platform (including user manuals) can be found at The DIstributed Computing Environments (DICE) team homepage at has information on projects which use Atmosphere for cloud resource provisioning

16 Cost optimization of applications on clouds
Infrastructure model Multiple compute and storage clouds Heterogeneous instance types Application model Bag of tasks Multi-level workflows Modeling with AMPL (A Modeling Language for Mathematical Programming) and CMPL Cost optimization under deadline constraints Mixed integer programming Bonmin, Cplex solvers M. Malawski, K. Figiela, J. Nabrzyski: Cost minimization for computational applications on hybrid cloud infrastructures, Future Generation Computer Systems, Volume 29, Issue 7, September 2013, Pages , ISSN X, Maciej Malawski, Kamil Figiela, Marian Bubak, Ewa Deelman, Jarek Nabrzyski: Cost Optimization of Execution of Multi-level Deadline-Constrained Scientific Workflows on Clouds. PPAM (1) 2013:

17 Simulation and scheduling of large-scale scientific workflows on IaaS clouds
Large-scale scientific workflows from Pegasus workflow management system Workflows of 100,000 tasks Workflow ensembles: schedule as many workflows as possible within a budget and deadline Cloud infrastructure simulated using CloudSim M. Malawski, G. Juve, E. Deelman, J. Nabrzyski: Cost- and deadline-constrained provisioning for scientific workflow ensembles in IaaS clouds. SC 2012: 22

18 Cloud performance evaluation
Performance of VM deployment times Virtualization overhead Evaluation of open source cloud stacks (Eucalyptus, OpenNebula, OpenStack) Survey of European public cloud providers Performance evaluation of top cloud providers (EC2, RackSpace, SoftLayer) A grant from Amazon has been obtained M. Bubak, M. Kasztelnik, M. Malawski, J. Meizner, P. Nowakowski and S. Varma: Evaluation of Cloud Providers for VPH Applications, poster at CCGrid th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, Delft, the Netherlands, May 13-16, 2013

19 DICE team - http://dice.cyfronet.pl
Main research interests: investigation of methods for building complex scientific collaborative applications and large-scale distributed computing infrastructures elaboration of environments and tools for e-Science development of knowledge-based approach to services, components, and their semantic composition and integration CrossGrid interactive compute- and data-intensive applications K-Wf Grid knowledge-based composition of grid workflow applications CoreGRID problem solving environments, programming models GREDIA grid platform for media and banking applications ViroLab GridSpace virtual laboratory PLGrid series advanced virtual laboratory, DataNet gSLM service level management for grids and clouds UrbanFlood Common Information Space for Early Warning Systems MAPPER computational strategies, software and services for distributed multiscale simulations VPH-Share federating cloud resources for development and execution of VPH compute- and data-intensive applications Collage Executable Papers; 1st award in the Elsevier Grand Challenge competition at ICCS2011 ISMOP cloud resource management and optimization, big data storage and analysis tools PaaSage federating cloud resources, workflow composition, optimization of cloud resources, porting existing applications to cloud infrastructure


Download ppt "SimCity Building Blocks at the DICE team"

Similar presentations


Ads by Google