International Symposium The University of Texas in Arlington, 28 - 30th July Part 1: Mathematical modeling of processes in systems life cycles in compliance with standards requirements of ISO/IEC 15288 Part 2: Methodical Approach for the Evaluation of System Vulnerability in Conditions of Terrorist Threats Prof. Kostogryzov Andrey, Russia, Moscow, akostogr@ceisoq.ru Plan (part 1) 1. The Role of Mathematical Modeling. Abstract Idea for Information Systems Modeling 2. Offered Models and Software Suites CEISOQ 3. Mathematical Modeling for Process Architectures Analysis, Examples 4. Mathematical Modeling of Items Gathering , Examples 5. Mathematical Modeling of Item Content Analysis , Examples 6. Mathematical Modeling of System Protecting , Examples 7. Mathematical Models in Developing , Examples 1.1 The Role of Mathematical Modeling. Abstract Idea for Information Systems Modeling LIFE CYCLE STAGES CONCEPT DEVELOPMENT PRODUCTION UTILIZATION SUPPORT RETIERMENT PURPOSE Identify stakeholders' needs Explore concepts Propose viable solutions Refine system requirements Create solution description Build system (including mathematical software models as a system component) Verify and validate system Produce individually or in quantity Inspect and test Operate system to satisfy users’ needs Provide sustained system capability Store, archive or dispose the system MODELS CAPABILITIES Substantiation of quantitative system requirements Requirements analysis, investigations of risks and potential threats, evaluations of proposed solutions and expected hazards. Tests and evaluations of system operation quality Investigations of risks and potential threats. Evaluations of system operation quality, optimization of parameters Substantiation of rational modes Practice solutions are based on system analysis: the fundamentation for providing high quality of system is rational use of models and methods that allow to evaluate and optimize different existing processes in life cycle 1.2 The Role of System Analysis and Mathematical Modeling - the main actions for rational achievement system purposes For example in Project Process: - for Project Assessment Process - assess project progress, analyze data and measures to make appropriate recommendations; - for Risk Management Process - evaluate the risks, determine the risk treatment strategies; - for Information Management Process - define information maintenance actions in Technical Processes: -for Requirements Analysis Process -define the functional boundary of the system, technical and quality in use measures, specify system requirements and functions; - for Architectural Design Process - define appropriate logical architectural designs, evaluate alternative design solutions; - for Operation Process - monitor the system operation, introduce remedial changes to operating procedures, the operator environment, humanmachine interfaces and operator training as appropriate when human error contributed to failure etc. 1.3 Abstract Idea for Information Systems Operation Modeling SYSTEM Higher systems Purposes Resources Sources Requirements to information system Use conditions Information system Operated objects Users Interacted systems The general purpose of operation: to meet requirements for providing reliable and timely producing complete, valid and confidential information for its following use Subordinate systems A totality of system characteristics that bears on ability to satisfy users needs in output information defines information systems operation quality 1.4 Abstract Idea for Information Systems Modeling Required information (ideal) Reliable, timely, complete, valid and confidential information Used information (real) non-confidential due to processing intolerable mistakes required quality doubtful non-produced as a result of system's unreliability untimely non-actual with hidden distortions as a result of unauthorized accesses with hidden software distortions due to random faults of staff and users incomplete due to random faults missed during checking 2 Offered Models and Software Suites CEISOQ CEISOQ was presented on conferences: “Modeling and Certifying of Arms and Military Techniques” (1998-2000), on the International Air-Space Show MAKS (1999, the town of Zhukovsky), on the International Engineering and Technical Management Symposium (2000, the USA), on the International Computation and Information Conference (2000, Kuwait), on the International Seminar “Mathematical Methods, Models and Architecture for Providing Computer Networks Security” (2001), was awarded be the Golden Medal of the International Innovation and Investment Salon (2001, the chairman of Jury is the Nobel Prize Winner Mr. Alfyorov) etc., is patented and certified. 2.1 Model of Functions Performance in Conditions of System Unreliability (physics) Cause 1 t1 t1+q Reliability is provided Cause 2 t2 Cause 3 t2+q t3 t Reliability is not provided - Failure-free system operation time (operative conditions); - Failure recovery time (inoperative conditions); - Required permanent time of reliable operation. An application allows to answer such questions as • what requirements to items (hardware/software units, staff or users) operation time between failures and to repair time are admissible? • what operation processes should be duplicated? • what about the reliability of a system architectural design etc. 2.2 The Models Complex of Calls Processing (physics for information system) Calls for technological operations Produced output data Users calls for output information producing BUFFER Users calls for input information filling SERVING SYSTEM Performed technological operations Transferred messages Felt data Users calls for messages transfer An application allows to answer the questions: -what processing units should be choosen from the producing point of view? -what calls flows and functional tasks may be main causes for bottle-necks? etc. Software tools CEISOQ allows to compare effectiveness of six dispatcher technologies: • technology 1 for a calls processing without priorities: in consecutive order for unitasking processing mode; in time-sharing order for multitasking processing mode; • priority technologies in consecutive processing order: – technology 2 for calls processing with relative priorities “first in - first out”; – technology 3 with absolute priorities; – technology 4 for calls batch processing with relative priorities; – technology 5 combined by technologies 2, 3 and 4. 2.3 The Model of Entering into System Current Data Concerning New Objects of Application Domain (physics) а) Processes of new objects and events (OE) appearance and information delivery: OE № 1 OE № 2 OE № 3 OE № 4 t message № 1 message № 2 - information completeness message № 3 message № 4 - information incompleteness b) Modeling queuing system M/G/ Real process: OE appearance Information delivery Unit 1 Storage Database Objects features Events parameters Unit 2 Formalization: Calls flow … ... Served calls Calls serving An application allows to answer such question as what productivity of preparation, transfer and input units preferred to provide information completeness? and others should be 2.4 Models of Items Gathering from Sources (physics) In difference from Model 2.3 evaluated item’s actuality characterizes updating processes DB DB DB t DB - Items actuality is provided updating moments; Items actuality is not provided items actuality maintenance after the last updating; - non-actuality for items - An application allows to answer the question as what productivity of preparation, transfer and input units and what gathering technologies should be preferred to provide items actuality? 2.5 Models of Items Gathering from Sources (mathematics) The probability of information actuality maintenance until its use moment (proved on the base of the limit theorem for regenerative processes) is: а) for the mode of source information delivery immediately after essential object current state change: 1 P B(t )[1C (t )]dt, 0 b) for updating mode regardless without dependence on object current state is changed or not (for example, when gathering is regulated): 1 [ 1 Q ( t )][ 1 C ( t ) dB ( )] dt , P 0 q0 where C(t) is the probability definition function (PDF) of time between neighboring essential real station changes, – is the mean time; B(t) is the PDF of time for information preparing, delivering, and inputting; Q(t) is the PDF of time interval between the neighboring updating, q is the mean time 2.6 Model of Items Analysis (physics on example of information checking) Initial information preparation Attention concentration is restorated Attention concentration is OK ... ... Check-up and fault correction Formalized checking of fault lack Real information Fault lack Input information Fault lack Fault lack Fault ... Fault lack information (fault detected) Fault ... Fault Fault lack Fault lack A man is the most valuable systems component An application allows to answer the questions: Is a checker able to reveal all existing errors? Moreover, is he able to commit no own errors? Is there needed software support for effective analysis? What about system faultlessness in real time operation? etc. 2.7 Models Complex of Dangerous Influences on a Protected System (physics on example of computer viruses influences ) Established regulating diagnostic Validation result of initial system The state able to operate Functional task performance of System resources state diagnostic System transformation to «pure» state Functional task performance in established mode Hidden or is «purity» visible Virus penetration Virus activation violation system operation Dangerous influences (for example from computer bags viruses or terrorists etc.) define high system vulnerability not only through a suddenness and incomprehensibility of their influences but also mainly on account of their insufficient studyness. An application allows answer the questions: · what about a danger of influences on a protected system and which safety technologies should be preferred for different environment scenarious? etc. 2.8 Model of an Unauthorized Access to System Resources (physics) The core of modeling is barriers overcoming as random processes 1-st barrier to be overcome during unauthorized access * . . . * * * * * * Unauthorized k-th barrier to be overcome during unauthorized access Violator * * * * * * * * * * * * Stored system resources * * * Unauthorized actions ... * * * * * * * * * * * not realized access is * * * * Unauthorized access is realized * An application allows to answer such questions: - what about the quantity of barriers? - what protection barriers decoding time is tolerated against unauthorized accesses? etc. 2.9 Model with due regard to resources objective value (physics on example of information confidentiality) 1-st barrier to be overcome during unauthorized access * . . . * * * * * Information confidentiality is * * k-th barrier to be overcome during unauthorized access Violator * * * * * * * Unauthorized actions ... * * * * * * * * Stored system resources * * maintained * Keep* one’s patience, * please! Just 36,6% * * * * * * Period*of objective * value * after * beginning ! The Model has the difference from Model of an Unauthorized Access in considering a period of objective value of kept resources. One allows to answer the questions: what about a probability of unauthorized access with this consideration? Is this period valueable? etc. 3 Mathematical Modeling for Process Architectures Analysis For example a system project shall implement the following activities: - define appropriate logical architectural designs - analyze the system functions identified in requirements analysis and allocate them to elements of system architecture - analyze the resulting architectural design to establish design criteria - determine which system requirements are allocated to operators for the most effective, efficient and reliable human-machine interaction - evaluate alternative design solutions, modeling them to a level of detail that permits comparison against the specifications expressed in the system requirements and the performance, costs, time scales and risks expressed in the stakeholder requirements - specify the selected physical design solution as an architectural design baseline in terms of its functions, performance, behavior, interfaces and unavoidable implementation constraints etc. Example 3.1. Let an Air Transport System be developed for making intercontinental flights (see Fig. D.1 from ISO/IEC 15288): Aircraft System (subsystem 1); Airport system (subsystem 2); Fuel distribution system (subsystem 3); Air traffic control system (subsystem 4); Ticketing system (subsystem 5). Let from users point of view the mean inoperable technical conditions for a hypothetical Air Transport Systems is equal to 2 hour per year. And it is admissible. It means that admissible availability is equal to 0.99977 (1-2hours/(365days 24hours)). System recovery time after an error equals to 30 minutes. It is required to evaluate the availability and probability of reliable Air Transport System operation during 1 day in considering equal reliability for all subsystems. Results: 1) for providing required availability the mean time between failures for one subsystem should be equal more than 1.3.years and the mean time between failures for whole system will be about 0.26 years; 2) for providing required reliability during one day the mean time between failures for one subsystem should be equal more than 61 years and for the system will be about 12.2 years- more in 47 times (!!!) Example 3.2. An information basis for efficient fighters use is a board radar complex. It weighs about 1% of the whole airplane and its cost varies in the range 10-20% of the whole airplane cost. One of the bottlenecks of any modern fighter is low reliability of radar stations (RS). Mean time between failures equals to 200 hours is still considered as admissible. For assigned admissible probability of reliable performance of a set task not less than 0.95 it is required to define the maximum continuous time of a board RS use. Solution. Let’s the mean time of RS repair after its failure is equal to 4 hours. The maximum mean time of continuous RS use mustn’t exceed 6 hours 15 minutes. It is enough for a certain military task fulfillment by a fighter. If the mean time between failures equals 200 hours the probability of reliable RS operation while a fighter fulfils a military task can’t exceed the level 0.97-0.98 (!!!). Example 3.3. If any Transport System (see Fig.D1 from ISO/IEC 15288) is very large there are hundreds or even thousands of staff members. To solve a given functional system problem there are required efforts of several specialists. Let a problem solution depend on joint but independent actions of 5 people. Let each of 4 specialists make 1 error a month and the 5th inexperienced person makes 1 error a day. System recovery time after an error equals to 30 minutes. It is required to evaluate faultlessness of such a group’s actions within a week. Results. The probability of faultless joint actions of the first 4 specialists within a 40-hours workweek equals to 0.8-0.82 but the low-quality work of the 5th member mocks the whole group work. The probability of faultless actions decreases to 0.02 (!!!). These results prove the importance of thorough specialists training because a man is the main system bottleneck. 4. Mathematical Modeling of Items Gathering Example 4.1. Military aviation effectiveness is caused by intermediate-range ballistic missiles use. Distant air enemy detection and identification is carried out by an onboard radar station (RS). ABOUT All locators are divided into 2 groups: with mechanical air surveillance (3-rd generation) and with electronic scanning (4-th generation). The unavoidable disadvantage of existing mechanical scanning systems is the impossibility of combining two modes: missile guidance on a target and surveillance. This means that sequentially surveying air RS sets current positions of several just detected targets. In case of fighting with nonmaneuvering low-speed targets (for example with cruise missiles) this method proves to be correct. When enemy maneuverability is high the achieved information actuality turns out to be insufficient. To destroy a maneuvering target its data should be fixed 5-10 times per second. The thing is that an aerial with mechanical surveillance may define new target’s data only in its next turn, i.e. in a second. In this case a pilot does not have any ability to survey air, what in fight conditions causes an indubitable loss. What about information quality for locators of 3-rd and 4-th generation in quantity? Example 4.1 (computation results) Input: i is the mean time between essential changes; i is the mean preparing time; i is the mean transfer time; I is the mean time of entering into a system. Di = D2 – it means that information is gathered without any dependencies on changes, qi is the mean time between updating. Output: Pact.i is the probability of information actuality on the moment of use. Case of fight with non-maneuvering (i=1-4 are RS with electronic scanning, i=5-6 – with mechanical one) Case of fight with high-maneuvering enemy targets Example 4.1 (summary of modeling) Computation results show that information gathering process architecture based on RS with mechanical scanning provides information actuality not less than 0.88-0.94. Achieved probability 0.88-0.94 may be considered as standard for effectiveness in fighting with non-maneuvering or low-speed targets. For comparison use of electronic scanning RS allows to increase this probability to 0.96-0.98. The difference in case of high-maneuvering enemy targets is in the frequency of significant targets positions changes – they happen every 1-2 seconds. Computation results prove that mechanical scanning RS may provide information actuality equal to 0.56-0.74. Electronic scanning RS may provide used information actuality equal to: 0.81 - 0. 85 in case of sharp turn maneuvers within a second; - 0.90 - 0.92 in case of turn maneuvers within 2 seconds. Conclusion: - actuality increase to the level 0.9 and higher is a very important and necessary condition for fighting aviation effective opposing of fighting aviation to high-maneuvering targets. Due to such high information actuality there is possible for a system to obtain a new quality (practice “transition from QUANTITY to QUALITY”) Example 4.2. Information roots in “Wall Mart” success are on the first place. To increase productivity each worker salesclerks were equipped with manual bar-code readers. Information contained in a bar-code is shown on a display. A worker can get a retrospective picture of products sold within a day, a week and 5 weeks. On each article there is everything what may be necessary for ordering. What may be achieved due to information actuality increase? Summary. Actuality of information is not less than 0.992 (i=1-4). Information read from a bar-code, which is transferred to the “Wal-Mart” headquarters.The satellite system is connected with more than 4000 company suppliers. For comparison, other shops, where information is updated hourly, information actuality equals to 0.3-0.7 (i=5-8), i.e. at a moment of information use it is as true as false. Feel the difference !!! 5. Mathematical Modeling of Item Content Analysis Example 5.1. Let an operators’ work order of an air traffic control system (see Fig.D1 from ISO/IEC 15288) be developed. As the main operator’s work is connected with monitors it is required to substantiate time of a continuous operator’s work shift under the condition that the probability of obtaining correct results of information analysis is not less than 0.99. Data mining for modeling Let’s the number of simultaneously tracked objects does not exceed 20. Changes of object states happen every second. During an hour of work there happens not more than 3600 changes of each object state. As a flight is continuous it is critical for an operator if the frequency of object state changes equals to 1 change per 5 seconds. Let’s assume that among these 20 objects there are not more than 4 significant ones preparing for a take-off or landing. Thus within an hour an operator must analyze up to 14400 objects’ states (3600/520), within 2 hours – 28800 ones, within 4 hours – 57600 ones, within 5 hours – 72000 ones. A fraction of essential information doesn’t exceed 20%. An analysis speed of an experienced analyzer equals 4 objects a second. The frequency of type I and II errors equals to 1 error a week. Example 5.1. (summary of modeling) Input: Vi is the content of analyzed information; i is the fraction of information essential for analysis; i is the information analysis speed; ni is the frequency of type I errors (when unimportant information is considered as essential); TMTBF i is the mean time between algorithmic type II errors; Tcont.i is the continuous period of an analyst’s work. Treq.i is the assigned term (deadline) for analysis. Output: P after i is the probability of correct analysis results obtaining; after i is the fraction of unaccounted essential information i=1-3 describe one-hour work and i=4-8 – five-hour work. Owing to high qualification of operators the computed correctness of their work is stably high – 0.94-0.98, but less than required one (0.99). Result: there are no formal solutions of the problem because the assigned probability is almost unachievable, the problem can’t be solved due only to the control of a shift’s work time Example 5.2. An automated monitoring system of a territorially distributed system critical to a safety problem is developed (a pipeline of oil or gas network, communications of a chemical enterprise, stores of nuclear industry waste, a regional energy system etc.). Information from automatic sensors is transferred to an integrating center. Though degree of control is quite high and control itself is continuous the main information gathering and processing functions are performed in an automatic mode. An operator receives integral information, making a decision and developing control actions. It is required to define such minimum speed of data processing that the probability of correct integrated results obtaining is not less than 0.9999. Result: only if data processing speed in the automatic mode is not less than 199500 symbols a minute and in the manual mode not less than 108 graphical results a minute there may be achieved the required correctness. The obtained results are system engineering requirements for a monitoring subsystem as well as requirements to qualifications of system staff. 6.1 Modeling of System Protecting From Unauthorized Accesses Example 6.1. In the end of 1941 allies belonging to the anti-Hitler coalition in the English naval began to form escorts with transport and military vessels, which were sent to northern ports of the USSR. In 1942 an escort PQ-17 was sent to the USSR. Suddenly on half the way the escort was attacked by German submarines and bombers. As a result 24 of 36 vessels were drowned, 3350 trucks, 430 tanks and 100000 tons of freight disappeared on bottom. The matter is the Finnish center of radio interception received a telegram in the Morse code, decoded one and transferred it to the Germans. It is required to substantiate system requirements to cipher complexity for providing latency of transition of a vessels caravan with the probability not less than 0.99. Result: to provide information confidentiality during 20 days it was required to select cipher algorithm for the cipher decoding more than 3 years (!!!) for complexity 6.1 Modeling of System Protecting From Unauthorized Accesses Example 6.2. Once on the competitions there was suggested to solve some problems connected with enciphering. There were compared an algorithm of 97-rate enciphering on the elliptic curve and an RSA algorithm with a 512-digits key. To take part in competitions there were got 195 enthusiasts from 20 countries. For deciphering, which took 40 days, there were used 740 computers. In the end deciphering took about 16 000 machine years on conversion to computers which throughput is 1million operations per second. It seems that the final results of the competitions are not of practical importance for a specialist (it was only revealed that a cipher may be deciphered in 40 days). But in fact the latent results are amazing not in the form they were drawn in competitions but in the form of cryptographic protection modeling results (!!!). Example 6.2. (system analysis) Let it is required to define maximum admissible period of objective information confidentiality that may characterize protected information to provide its confidentiality with the probability not less than 0.995. If the key is changed 1 time a year the enciphering will provide confidentiality for information, which period of objective confidentiality equals to more than 200 million years (!) in case of 97-rate encoding on the elliptic curve and not more than 4-5 months (!) in case of enciphering with the help of the RSA algorithms. If a key is changed once a month (practice for today) the encoding by algorithm RSA provides information confidentiality not less than 0.999, and the period of objective information confidentiality will also be equal to hundreds of millions of years!!! How may be estimated the following information??? “…The RSA Data Security firm recommends companies to protect data by more reliable keys, which length exceeds 768 digits or better 1028 digits…” Brilliant answer (ancient): effective system engineering decision should be made only on thorough modeling knowledge for system processes 6.2 Mathematical Modeling of System Protecting Against Dangerous Influences Technology 1 – preventive diagnostic of system integrity Technology 2 - multishift security monitoring Technology 3 - security monitoring when system integrity Input for modeling : j is the frequency of influences for penetrating a danger source; j is the mean activation time of a penetrated danger source (only for technologies 1 and 3); T betw.j is the time between the end of diagnostic and the beginning of the next one; Tdiag.j is the diagnostic time including the time of system integrity repair; TMTBF j is the mean time between operator’s errors; Treq.j is the required period of permanent secure system operation. Output: Pinf.j is the probability of dangerous influence absence in a system within the assigned period Treq.j; Ppen.j is the probability of penetrated danger source absence. Example 6.3. On the market has appeared a special medical system with sensors built in a man’s clothes, which immediately give warnings of dangerous changes of a man’s physical state. A special portable computer gathers information from sensors and processes it. Let’s evaluate what lifetime without any serious illnesses may be provided for people using such clothes. System analysis. Let’s examine 2 variants of a man’s reaction upon signals. The 1-st variant implies a visit to a doctor and taking the necessary treatment within 4 hours after a man has received a signal of danger from the system. The 2-nd variant implies an immediate intervention of a personal doctor after first danger symptoms appeared and an organism integrity recovery(reflected ). Results: the probability of serious illnesses absence is for 1-st variant - within a year not les than 0.98, within 2 years – not less than 0.92, within 10 years – not more than 0.35; for the 2-nd variant - within 46 years not less than 0.95. 7. Mathematical Models in Developing In compliance with standards requirements of ISO/IEC 15288 there are developed: - Model for enterprise environment development; - Models of project development; - Models of systems development in life cycle; - Models of life cycle process development; - Models of customer satisfaction in system life cycle and others The models may be applied for solving such system problems appearing in a systems life cycle as: . - substantiation of quantitative system requirements to hardware, software, users, staff, technologies; - requirements analysis; - evaluation of project engineering decisions and possible danger; - detection of bottle-necks; - investigation of problems concerning potential threats to system operation and information security; - testing, verification and validation of system operation quality; - rational optimization of system technological parameters; - substantiation of plans, projects and directions for effective system utilization, improvement and development Publications and Implementations 1994 1996 1999 2001-2003 Applications Summary: Presented models and software tools are the methodological and implementation foundation of customers, enterprises, certification bodies, test laboratories and experts in Russia. They support the Russian standard “GOST RV. Information technology. Set of standards for automated system. The typical requirements and metrics of information systems operation quality. General principles”. CEISOQ has already found wide application in Universities for education Plan (Part 2) 1. Approach for Evaluating System Vulnerability in Conditions of Terrorist Threats. General Propositions 2. Modeling complex and software suites “VULNERABILITY” 3. Investigations after September 11 3.1 How effective has been the system of flights safety provision before September 11 (in Russia and the USA) ? 3.2 How the level of the existing safety may be increased and by what measures? 4. Examples for a sea gas-and-oil producing system 4.1 Risk to become an object of terror 4.2 Risk of suspicious events non-detection and mistaken analytical conclusions 4.3 Risk of latent penetration and influence 4.4 Risk of protection barriers overcoming Conclusion 1.1 Approach for Evaluating System Vulnerability in Conditions of Terrorist Threats. General Propositions General scheme of terrorist threats development 1. Choice of a terrorist object 2. Search of system vulnerability 3. Preparation for terrorist acts 4. Latent or obvious actions concerning system vulnerability 7.** Analysis of terrorist threats realization results 6. Realization of threats by means of system protection barriers overcoming 5*. Formulation of requirements and threats * There may be no clear requirements and threats (as it was on September, 11 2001) ** For a single terrorist act there may be no feedback as a result of the terrorist impact analysis 1.2 Approach for Evaluating System Vulnerability in Conditions of Terrorist Threats. General Propositions Сhain of logical dependences for an evaluation of a system vulnerability risk 1.Risk to become an object of terror 2. Risk of suspicious events nondetection during admissible time 3.Risk of mistaken analytical conclusions . during admissible time 4.Risk of dangerous influence on system during given period System vulnerability risk during given period 5.Risk of protection barriers overcoming during given period 2 Modeling complex and software suites “VULNERABILITY” Rvulner - integral system vulnerability during given period tgiven (Psafety - integral system safety during given period tgiven), depends on: Pobject - risk to become an object of terror, Rnon-det - risk of suspicious events non-detection during admissible time, Rmist - risk of mistaken analytical conclusions during admissible time, Rinfl - risk of dangerous influence on system during given period, Rover- risk of protection barriers overcoming during given period 3.1 How effective has been the system of flights safety provision before September 11 (in Russia and the USA)? System analysis. Barriers from the point of view of terrorists: the 1st barrier is pass and interobject modes in aerodromes and centers of air traffic control; the 2nd barrier is a preflight examination and control of passengers and their luggage during the registration; the 3rd barrier is a preflight examination before boarding; the 4th barrier is a lock-up door to the cockpit; the 5th barrier is an on-line warning about a highjacking Input in case of opposing to inexperienced terrorists in Russia The computation results 3.1 How effective has been the system of flights safety provision before September 11? (computation results for trained terrorists ) in Russia in the USA Conclusion: In Russia and the USA the existing before September 11 systems of flights safety were uneffective against planned actions of prepared terrorists (protection in probability measure - about 0.52-0.53). The bottlenecks were a weak protection of a cockpit and absence of active opposing measures on board an airplane 3.2 How the level of the existing safety may be increased and by what measures? The first measure consists in making a cockpit door a real barrier insuperable for terrorists during a flight. As soon as a cockpit door has become insuperable a cockpit may be turned into a center of a cabin security telemonitoring and control. It is the second measure. Thus pilots may timely get complete and valid information about situation in the cabin. Terrorists who have revealed themselves are in standing positions; the others remain sitting. The first task of the defended part is to disable these threats subjects at least for a few minutes. There are needed means and ways of illethal action, i.e. which don’t lead to a fatal end because passengers may also run the danger. The third measure is a use of point means of illethal action on the revealed terrorists. Such means may be a soporific gas and/or short-period influences, for example, dazzling (a terrorist is suddenly dazzled by a light beam), and/or deafening, and/or electro-shocking causing a temporary loss of consciousness. There should be several ways of influence because one way may be easily neutralized (gas masks against gases, sunglasses against dazzling etc). Thus the revealed terrorists may certainly be disabled. As in a cabin may be accomplices able to repeat the high-jacking after an additional preparing there must be provided ways of compulsory keeping of suspicious passengers on their seats till the emergency landing. It is the fourth measure, which may be provided again by soporific actions, jammed belts etc. 3.2 How the level of the existing safety may be increased and by what measures? (computation results) Result. Owing to active opposing measures undertaking on board an airliner the probability of flights safety provision against terrorists may essentially increase from 0.52-0.53 to 0.98-0.99. Note: the first failures will make terrorists analyze their causes and find new bottlenecks of the security system thus continuing the counteraction. And preventive actions should base on modeling 4. Examples for a sea gas-and-oil producing system . Processes of possible terrorist influence and system safety provision (platforms, coastal technological complexes including floating storage and offloading terminals, liquefied natural gas terminals, pipelines, tubing stations) are evaluated for various scenarios and conditions 4.1 Risk to become an object of terror Let the mean time of revealing interesting information by terrorists be two days, the mean time of delivering the interesting information to the terrorist analytical center be no longer than 1 hour and the mean time of making a decision concerning the system attractiveness from the point of view of its possible choice as an object of terror be about one month. It is assumed that information gathering doesn’t obviously depend on system state changes and the mean time between information updates in the terrorist analytical center is forecasted about 1 month. There is an objective danger and high risk for sea gasand-oil producing systems to become an object of terror 4. 2 Risk of suspicious events non-detection and mistaken analytical conclusions Is it principially possible to solve the problem of detecting suspicious events in time and making correct conclusions from the gathered on-line information? Results: The analysis carried out on the basis of the facts concerning FBI activity has shown that the risk of erroneous analytical conclusions and as a consequence non-undertaking or undertaking inadequate countermeasures is equal more than 0.998 (!!!). Nowadays it is practically impossible to prevent realization of terrorist acts aimed at any kinds of systems and objects. It is necessary to conduct a profound purposeful work (based on modeling) on radical reduction of risks concerning non-detecting of suspicious events, erroneous analytical conclusions 4.3 Risk of latent penetration and influence For existing safety systems the probability that dangerous influences (explosions of fuel-air mixes clouds, generation and burning of fire balls, oil spill and burning, separation and spread of technological equipment parts and others ) doesn’t occur within 24 hours is above 0.99997 , i.e. the risk of emergency realization is about 0.00003. In conditions of daily failure danger the risk of required safety within a month increases up to 0.001. This high level of sea gas-and-oil producing system protection in emergencies is mainly provided by application of approved automatic safety technologies (!) For similar conditions the risks of terrorist threat realization are incommensurably higher (within a month risk runs up to 0.93). Owing to insufficient preparedness and technical equipment of operators for timely and valid recognition of terrorist threats at the background of other technical threats variety sea gas-and-oil producing systems are today completely helpless in case of terrorist dangers (!!!) 4.4 Risk of protection barriers overcoming In practice the protection from an unauthorized access is a sequence of barriers after successful overcoming of which a violator can get an access to system’s resources. Such an access is possible from a special command post when an automated workstation of control service is used. Change frequency of the barrier parameter value Mean time of barrier overcoming 1. Protection by a patrol boat Change of guards every 24 hours 10 hours Latent penetration from the air, under the water, fraud of guards 2. System of passes to the platform with a change of security services Change of guards every 24 hours 10 minutes Documents falsification, conspiracy, fraud 3. the electronic key to get to the GOPS control unit 5 years (time between changes) 1 week Theft, forcible key withdrawal, conspiracy 4. The password to enter the automated GOPS system 1 month. 10 days Spying, compulsory questioning, conspiracy, selection of a password 5. The password to get access to software devices 1 month. 10 days Barrier Possible way of barrier overcoming — — 6. The password to get access to the required information 1 month. 10 days 7. The registered external information carrier with write access 1 year 24 hours — — Theft, forced registration, conspiracy 8. Confirmation of a user identity, during a session of work with the computer 9. Telemonitoring of a helipad, a drilling module, energy equipment, a technological module, pipelines and equipment, rescue rooms and boats etc. 10. Encoding of the most important information 1 month. 24 hours Spying, compulsory questioning, conspiracy Time between changes of software devices) 1 month Simulation of a failure, false films, dressing up as employees, conspiracy Change of keys every month 1 year Decoding, conspiracy 4.4 Risk of protection barriers overcoming 1. The first 3 barriers are overcome with the probability about 0.34. Use of alternating passwords once a month for the 4th, 5thand 6th barriers allows to increase the protection level from 0.66 to 0.86. 2. Introduction of the 7th and 8th barrier is practically useless. 3. Use of telemonitoring means allows to increase information resource protection from UA to 0.998 what also doesn’t meet the stated requirements. 4. Only the use of all 10 barriers provides the required protection from UA with the required probability more than 0.9999 what practically excludes any risk of terrorist access to system control post information resources and essentially reduces system vulnerability. 4.5 Summary result for sea gas-and-oil producing systems Protection technologies from terrorists threats are imperfect and cannot be compensated by approved existing safety technologies. A whole system and its components (firstly platforms, coastal complexes including floating storage and offloading terminals, liquefied natural gas terminals, pipelines, tubing stations) are in fact extremely vulnerable (!) CONCLUSION Engineering decisions concerning development of protection system for preventing terrorist threats should be modeled, quantitatively evaluated and substantiated taking into account potential scenarios of threats development, opportunities of protective barrier overcoming during possible acts of terrorism etc. The offered complex “Vulnerability” is capable for system mathematical modeling, risks analysis and protection measures effectiveness investigations. A good beginning (based on modeling) is half the battle…