TOSCA Monitoring Straw-man for Initial Minimal Monitoring Use Case Roger Dev CA Technologies Revision 3 May 21, 2015.

Slides:



Advertisements
Similar presentations
What is proper format for the XDW document. In its first year, XDW has been exposed to feedback, and this public comment phase –to allow clarifications.
Advertisements

Mapping Service Templates to Concrete Network Semantics Some Ideas.
What’s coming in Sccm 2007R2 aka Sccm 2007R2: 10 reasons to upgrade Kim Oppalfens SCUG.be.
This product includes material developed by the Globus Project ( Introduction to Grid Services and GT3.
TOSCA Monitoring Straw-man for Initial Minimal Monitoring Use Case Roger Dev CA Technologies April 24, 2015.
DataGrid is a project funded by the European Union 22 September 2003 – n° 1 EDG WP4 Fabric Management: Fabric Monitoring and Fault Tolerance
Chapter 19: Network Management Business Data Communications, 4e.
The Architecture Design Process
1 ITC242 – Introduction to Data Communications Week 12 Topic 18 Chapter 19 Network Management.
Wireless LAN Topology Visualiser Project Supervisor: Dr Arkady Zaslavsky Project Team Members: Jignesh Rambhia Robert Mark Bram Tejas Magia.
Grid Computing, B. Wilkinson, 20046c.1 Globus III - Information Services.
Polaris Financial Technologies Welcomes the members of Hyderabad chapter for the 2nd event on 4 th July 14 held by PACE (The Testing Practice)
Web Service Architecture Part I- Overview and Models (based on W3C Working Group Note Frank.
1 Doctor Fault Management 18 May 2015 Ryota Mibu, NEC.
System Design/Implementation and Support for Build 2 PDS Management Council Face-to-Face Mountain View, CA Nov 30 - Dec 1, 2011 Sean Hardman.
SNMP ( Simple Network Management Protocol ) based Network Management.
BMC Software confidential. BMC Performance Manager Will Brown.
TOSCA Monitoring Working Group Status Roger Dev August 10, 2015.
1 Autonomic Computing An Introduction Guenter Kickinger.
Wave Relay System and General Project Details. Wave Relay System Provides seamless multi-hop connectivity Operates at layer 2 of networking stack Seamless.
An Introduction to Software Architecture
An Introduction to IBM Systems Director
SOS EGEE ‘06 GGF Security Auditing Service: Draft Architecture Brian Tierney Dan Gunter Lawrence Berkeley National Laboratory Marty Humphrey University.
© 2009 Voltaire Inc.1 Fabric Management in VM environment Marina Lipshteyn, Voltaire.
TOSCA Monitoring Working Group Status Roger Dev June 17, 2015.
Faculty of Computer & Information Software Engineering Third year
Computer Emergency Notification System (CENS)
(Business) Process Centric Exchanges
TOSCA Monitoring Reference Architecture Straw-man Roger Dev CA Technologies March 18, 2015 PRELIMINARY.
Device Identification & Driver Management TSC Update January 8, 2015.
Service Service metadata what Service is who responsible for service constraints service creation service maintenance service deployment rules rules processing.
Objective: Enable portability and semi-automatic management of applications across clouds regardless of provider platform or infrastructure thus expanding.
A Software Framework for Distributed Services Michael M. McKerns and Michael A.G. Aivazis California Institute of Technology, Pasadena, CA Introduction.
1 Service Creation, Advertisement and Discovery Including caCORE SDK and ISO21090 William Stephens Operations Manager caGrid Knowledge Center February.
Security Vulnerabilities in A Virtual Environment
Chapter 4 Version 1 Virtual LANs. Introduction By default, switches forward broadcasts, this means that all segments connected to a switch are in one.
Foundations of Information Systems in Business. System ® System  A system is an interrelated set of business procedures used within one business unit.
Ceilometer + Gnocchi + Aodh Architecture
MTA EXAM Software Testing Fundamentals : OBJECTIVE 6 Automate Software Testing.
Network management Network management refers to the activities, methods, procedures, and tools that pertain to the operation, administration, maintenance,
Role Of Network IDS in Network Perimeter Defense.
Slide 1 2/22/2016 Policy-Based Management With SNMP SNMPCONF Working Group - Interim Meeting May 2000 Jon Saperia.
TOSCA Monitoring Straw-man for Initial Minimal Monitoring Use Case Roger Dev CA Technologies Revision 2 May 1, 2015.
Keith Chadwick 1 Metric Analysis and Correlation Service. CD Seminar.
I2rs Requirements for NETCONF IETF 93. Requirement Documents
Failure Inspection in Doctor utilizing Vitrage and Congress
In Depth Azure StackIn Depth Azure Stack Resource Providers Damian Flynn MVP Daniel Savage Microsoft.
READ ME FIRST Use this template to create your Partner datasheet for Azure Stack Foundation. The intent is that this document can be saved to PDF and provided.
SDN-O LCM for Mercury Release Key Points and Overview
SQL Database Management
ONAP and MEF LSO External API Framework Functional Reference Architecture 12 July 2017 Andy Mayer, Ph.D. © 2016 AT&T Intellectual Property. All rights.
Chapter 19: Network Management
Fault Management with OpenStack Congress and Vitrage, Based on OPNFV Doctor Framework Barcelona 2016 Ryota Mibu NEC Ohad Shamir Nokia Masahito Muroi.
LOCO Extract – Transform - Load
Cisco Data Virtualization
Cloud Management Mechanisms
LSO Hackathon Summary Charles Eckel, Cisco DevNet.
Deploying and Configuring SSIS Packages
Network Administration CNET-443
Cloud Application Marketplaces
20409A 7: Installing and Configuring System Center 2012 R2 Virtual Machine Manager Module 7 Installing and Configuring System Center 2012 R2 Virtual.
Cloud Application Marketplaces
SUPA Policy-based Management Framework (SUPA: Simplified Use of Policy Abstractions) draft-ietf-supa-policy-based-management-framework-01 Will Liu, John.
Cloud computing mechanisms
An Introduction to Software Architecture
SNMP (Simple Network Management Protocol) based Network Management
PLANNING A SECURE BASELINE INSTALLATION
Cloud Application Marketplaces
Caller ID for Managed Critical Communication
Presentation transcript:

TOSCA Monitoring Straw-man for Initial Minimal Monitoring Use Case Roger Dev CA Technologies Revision 3 May 21, 2015

TOSCA Monitoring Use Cases (full – From Arch Ref Strawman) As an Application Architect, I want to know which metrics I can expect to be available from a given component. As an Application Architect, I want to define, within the Service Template, the Metrics to be collected for a component, as well as how they are to be collected and managed (thresholded, etc). – Additionally, I may want to use my favorite monitoring tools rather than those provided by the Service Provider. As an Application Operator, I want to be able to access collected metrics and events for any or all of my deployed components, either: – Interactively – Programmatically As an Application Developer, I want to be able to produce custom metrics from my application(s) and have them stored and accessed along with any standard metrics As a Service Provider, I may want to define Monitoring Policies for component types that may be different from those designated by the Application Architect. As a Service Provider, I want to be able to utilize my favorite monitoring tools rather than those supplied with an orchestration framework. As a Service Provider, I want to be able to access a robust set of Metrics and Events about the orchestration framework, since that is a critical component of my infrastructure. As a Service Provider, I want to be able to utilize the full set of topological information provided by the Nested Service Template(s) to enhance my knowledge of running applications. This includes the output sections of the Templates.

Revised approach based on feedback and discussion to date Simplify the initial use-case to the bare minimum Use that to work through the basic mechanisms Define and agree upon the fundamental mechanisms Expand from there

TOSCA Monitoring Reference Diagram Monitoring Automation Point (MAP) MAD (Monitoring Act / De-act) MIA (Monitorning Info Access) MEA (Monitoring Extension Advert) OM (Orchestrator Monitoring) Service TemplateExternal ProcessInternal Process External Monitoring System - Monitoring Template / Policy -Management Communication Info - Metric Availability - Metric Time Series - Events? -Metric Values -Events? -Metric Time Series -Events Focus on Subset of MAD

Initial Minimal Use Case (1 of 2) Assume we have a mechanism for defining the metrics associated with a given component-type: – This is a tractable problem, so let’s come back to it after solving more fundamental issues Assume that we want to monitor all the metrics that each component can produce – Defer defining the mechanism whereby the Application Architect can define the monitoring policy – Defer defining the finer points of policy (e.g. events, actions, transformations, etc.)

Initial Minimal Use Case (2 of 2) Scenario: – A Service Template is deployed with a single SoftwareComponent running on a single ComputeNode (virtual). Metrics (Capabilities) are defined for both SoftwareComponent and ComputeNode types. – ComputeNode: » PercentCpuUtilization, IoBytesIn, IoBytesOut – SoftwareComponent: » PercentCpuUtilization, TransactionsProcessed, ErrorsEncountered – Metrics are collected for some time – The Service Template is removed (de-deployed)

Minimal Use Case Constraints The components themselves (e.g. the ComputeNode -- via virtualization software, and the SoftwareComponent) will not be required to implement a new monitoring protocol – One of several existing monitoring protocols could be used (SNMP, WMI, Proprietary,etc.) depending on the service provider and the underlying technologies used The Monitoring Sub-System (MSS) is not required to be embedded within the Orchestrator. – An off-the-shelf monitoring system may be employed The MSS is running and attached to the Orchestrator before the Service Template is deployed

Notes on Monitoring Agent We should consider that there is always a Monitoring Agent (or give it a new name) – So called “Agentless Monitoring” just means that the Monitoring Agent role is baked into the component and doesn’t have to be explicitly added in. – From the MSS side, the only difference is the particular protocol used, and elimination of the need for an agent deployment step. Coordination is, in any case, still needed between the MSS and the Agent role within the component (address, port, creds, and other identifiers). In many cases, the Agent capability of one component is used to monitor a different component (e.g. one might use the hypervisor’s Agent to monitor a VM; one might use the Host OS’s agent to monitor an application process)

Diagram for Scenario 1 Service Template Virtual Machine (new) Software Component (new) Causes HostedOn Monitoring Sub-System (MSS) Notify State Change: -Create -Modify -Destroy Deliver Metrics (any existing push or pull protocol) 1 2 3

Notes for Scenario 1 Diagram New components are created. In some cases, there must be relationship information for components that are created outside of the ST (such as hypervisor or physical system -- see 2 below) MSS is notified of new components to be monitored. MSS Needs: – Service Template meta-data in order to know the ID and Type of the component – Instance Model in order to know the address, port, and credentials needed in order to collect metrics – Possibly the relationship to components not in the ST (e.g., the hypervisor) if info about the component is provided by that outside component. If push protocol, the monitoring agent, within the component, must be configured with the address of the MAP, and the TOSCA id of the component. If pull, then there must be coordination between Orchestrator and the monitoring agent, or explicitly defined in the ST, so that the create notification can know e.g., the agent’s port address and creds

Minimal Scenario Questions to Answer What information is needed by the Monitoring Sub-System (MSS) in order to activate monitoring when the Service Template is deployed. What mechanisms could be used to notify the MSS of the significant state changes for the components? – Activate – Modify – Deactivate What is the simplest mechanism that could handle this scenario

What information does MSS need? Agent Address and Protocol (How to talk to agent) Component Identifier (How to ask the question about the correct component) Credentials to access Agent – Might be able to set up a closed management network and not need creds???

Potential Mechanisms M1 – When a component is activated, make correlated Template Model and Instance Model available to MSS. MSS figures out how to monitor based on this info: – Proximate portions of Models are extracted and passed or an interactive API for browsing relationships is provided – Assume that Agents are either baked into the components, or are explicitly deployed by Service Template. Others???

What do we need to specify? What metric available for each Component Type? – Metric Type ID – Description – Data Type (e.g. Numeric, String, etc.) – Units (e.g. Volts, Megabytes, Percent, etc.) – Constraints (Min, Max, Enumerated Values, etc) Monitoring Policy (controlled via Service Template): – Monitoring Disposition: Required – Don’t deploy if you can’t monitor Best Effort – Deploy anyhow but enable monitoring if available None – Metrics to Include? Exclude? – Components to Include? Exclude? – Minimum Sample Frequency? – Action Conditions (e.g. If A then Do B) Not in this phase?

Metric Types Availability: – Percent Available Performance: – CPU Usage – I/O – Memory Workload: – Units Processed – Failed Units – Bytes Processed Security – Access Attempts – Failed Access Attempts Locally Defined?