Listing all articles in The Residual World under the category 'Architecture Framework' :

Definitions - What Exactly is a Risk?

by Nic Plum on Tuesday 12 March, 2013 - 22:30 GMT

Posted in Architecture FrameworkTRAKStandards

Tags: defencedefinitiondodiecnistsafetysecuritystandardtrakusa

NIST logoIEC logoUS DoD logo

Creating a definition sounds as thought it ought to be easy. It isn’t for many reasons - some of these are not so much technical as the process by which consensus is reached and the need to get consensus. For example the need to get consensus might mean that at times a weaker definition escapes because it was too difficult to get consensus with a tighter one.

Why do we care? Well there is a particular and a more general reason. The more general one is that the graphic blocks we use to represent the real world things have definitions and therefore the architect is supposed to select the most appropriate block to represent the real world thing based on the description. We can’t just choose anything otherwise we end up “head-modelling” where the verbal description we provide is not supported by the semantics of the model we’ve created (the model in our head is not the one on paper). If the description is wrong it might not be the right block to use (you wouldn’t represent ‘tank’ with a ‘tree’).

The particular reason is that we’ve a working group in TRAK looking to see if and how it is possible to extend TRAK to enable it to be used to address typical safety-related and security-related concerns. One of the starting points is therefore a review of general literature and particularly standards to identify the potential concepts or entities likely to be needed. In doing so we’ve found some potential problems with definitions.

A candidate entity is risk. What is a risk?

IEC 61508:2010

combination of the probability of occurrence of harm and the severity of that harm

MIL STD 882E

Mishap Risk. An expression of the impact and possibility of a mishap in terms of potential mishap severity and probability of occurrence.

NIST

The net mission/business impact considering (1) the likelihood that a particular threat source will exploit, or trigger, a particular information system vulnerability and (2) the resulting impact if this should occur.

There is a common thread. Many other standards also have very similar forms of definition. None of these, however, defines what a risk actually is The analogy is defining force as the product of mass and acceleration - it tells us nothing of what force is. None of the above are therefore definitions of risk they just indicate how we might derive a metric for it. One of the principles in defining something has to be that the definition is independent of other variables or an implementation. In the above if risk didn’t involve probability of occurrence it would mean that the concept of risk itself had changed which isn’t true.

My dictionary provides:

a possibility of harm or damage

IEC 61508:2010 defines a Hazard:

potential source of harm [Guide 51 ISO/IEC:1990]..

’ which is fine but then in the note that follows it states ‘….for example, release of a toxic substance…’ which looks to be a hazardous event not a hazard.

All of this means that it is harder and takes longer than it should do to analyse and form a view of a pragmatic compromise because you have to examine every word and be selective in what you choose to accept and what you choose to reject. You cannot blindly assume that any standard is correct since it is as much the product of gaining consensus as it is the technical content. You have to be a skeptical enquirer and constantly challenge. Too often folks put such committees on pedestals and don’t stop and think.

Comments

Comment on this article

Risk and Threats - The Common Ground Between Security and Safety?

by Nic Plum on Tuesday 10 April, 2012 - 21:25 GMT

Posted in Architecture FrameworkTRAK

Tags: def standefenceforumiso42010mil stdontologyrisksafetysecuritysolutionsourceforgestandardthreattrakviewviewpointvulnerability

TRAK Logo

This is something that has been bumbling around for some considerable time - safety and security. By that I whether there is something useful that an enterprise architecture view can be used for in the system safety and security disciplines.

On the face of it there is quite a bit of overlap. Both are ultimately concerned with risk inherent in a solution design which arises from threats (security) or hazards (safety). Both involve management with the aim to reduce the risk, threat or accident (safety) to an acceptable or tolerable target. I suspect also that security management also uses categories to classify acceptable severity or probability in much the same way that the various system safety management standards in defence do (MIL STD 882D, DEF STAN 00-56). Both also involve mitigation of risk by design - through structure, behaviour, or adherence to a normative process of some sort.

There are bound to be some differences, not the least of which is terminology. In the security area we seem to have constructs like:

  • Threat poses Risk
  • Threat exploits Vulnerability
  • design aka TRAK:Resource (System, Software, Organisation, Job or Role) exposed to Risk (and subsequently that Risk is mitigated by the (improved) Resource or Function (of that Resource)

In the safety area we seem to have constructs like:

  • Failure may present Hazard
  • Hazard can cause Accident
  • Accident poses Risk
  • Resource exhibits Failure

and attributes such as probability, impact, severity.

Anyway it seems sensible to open up the debate so I’ve posted some thoughts in the forums within the TRAK Viewpoints project site on Sourceforge. Something is definitely needed and my hunch is that there is so much overlap that it would be possible to create a Viewpoint that addresses the risk within a solution design. This may of course end up being two viewpoints depending on the concerns and therefore concepts (metamodel stereotypes) and relationships involved. What is needed is more debate and input from those involved with system safety and system security - hence the post. As ever with TRAK the objective is economy so that we have something that is just or barely adequate to describe the concerns and concepts involved and no more.

 

Comments

Comment on this article

Forums

Every Viewpoint Has to Be Distinct - Say “Goodbye” to the TRAK CVp-02 Concept Viewpoint

by Nic Plum on Sunday 08 April, 2012 - 12:42 GMT

Posted in Architecture FrameworkTRAK

Tags: concepttrakviewpoint

TRAK Logo

Every viewpoint in TRAK is a specification for an architecture description view. In accordance with ISO/IEC 42010 each address one or more typical concerns using a combination of tuples (stereotype - relationship - stereotype combination taken from the TRAK metamodel). The tuples have therefore to contain the right types and relationships to address the concern and the concerns (and therefore the tuple sets) must be distinct from those addressed by other viewpoints. This keeps clear water between viewpoints and it means that the number of viewpoints needed is kept to a minimum because they aren’t driven by domain or application of viewpoints.

What then of the TRAK CVp-02 Concept Viewpoint? This is currently defined as answering concerns has the concept purpose been identified? and How is it seen as being used? and the tuples as:

Expected to be largely textual and scenario based but with use of other concept perspective architecture views to illustrate, expand, define.The set of tuples will be those from the mandatory sets of the concept perspective views used against CVp-01, CVp-03, CVp-04, CVp-05 and CVp-06.The selection of concept views used to illustrate the scenarios is left to the architect.

from TRAK. Enterprise Architecture Framework Viewpoints. 2nd October 2011

This isn’t good enough. None of this needs anything which isn’t already provided by one or more of the other viewpoints in the TRAK Concept Perspective. The purpose of a concept is embodied through its relationships with the solution or potential solutions that realise it and its relationship with the enterprise and the enterprise goals. The content of a concept is already covered by existing viewpoints and there is nothing that makes this viewpoint distinct from any others. Historically it was an analogue of the MODAF OV-1 which included a high level graphic and a textual version used to present ideas to senior management in an easy to digest form:

The OV-1a provides a graphical executive summary of the architectural endeavour, which describes the interactions between the subject architecture and its environment, and between the architecture and external systems. A textual description accompanying the graphic is essential, with labels on the graphic and a detailed description in the OV-1b. Graphics alone are not sufficient for capturing the necessary architecture data.
The purpose of OV-1a is to provide a quick, high-level description of the business objective that the architecture is addressing, and how that objective might be achieved. An OV-1a can be used to orient and focus detailed discussions. Its main utility is to communicate the purpose of the architecture to non-technical, high-level decision makers.

from The MODAF Operational Viewpoint. 26th April 2010.

In TRAK any view can be presented using graphical elements as long as the type of object is shown and with simple text labels on relationships it is easy to produce something that most people can simply read in a natural way so the presentation is never a justification for a separate viewpoint.

On the face of it there is no good reason for keeping this viewpoint and the best thing is to remove it. The recommendation has been made as a change request (#3475115) and unless anyone makes a good reason to keep it the sentence will soon be carried out ....

 

Comments

Comment on this article

All articles/posts © of the respective authors

Site design and architecture © 2010 - 2011 Eclectica Systems Ltd.