logseq/pages/hls__JSSOFTWARE-D-22-00977_reviewer.md

file-path:: file://C:/Users/david/Zotero/storage/CBP56TTH/JSSOFTWARE-D-22-00977_reviewer.pdf

• Runtime monitoring ls-type:: annotation hl-page:: 2 hl-color:: green id:: 63b59e92-3ad2-42cb-94b7-c3d93d0d9235 hl-stamp:: 1672846996863
• reating probes to instrument the system, and defining constraints to be checked at runtime ls-type:: annotation hl-page:: 2 hl-color:: green id:: 63b59ec1-e257-4f69-b01e-07f8193a10a5
• generic monitoring platforms do not adequately cover domain-specific monitoring requirements ls-type:: annotation hl-page:: 2 hl-color:: green id:: 63b59ecb-f5b8-4703-8193-76afe7140b3a
• Generating CPS Runtime Monitors ls-type:: annotation hl-page:: 2 hl-color:: green id:: 63b59ede-98fd-4f0c-bb07-714729f53973
• esults show that GRuM facilitates the creation and evolution of a runtime monitoring platform with little effort and that the platform can handle a substantial amount of events and data. ls-type:: annotation hl-page:: 2 hl-color:: green id:: 63b59f13-a778-48f7-ac2f-eb660328dee8
• we present GRuM(Generating CPS Runtime Monitors), a framework that combines model-driven techniques and runtime monitoring, to automatically generate a customized monitoring platform for a given SuM. ls-type:: annotation hl-page:: 2 hl-color:: green id:: 63b59f6e-a0c7-44cb-a21e-5ad1da709790
• nmanned Aerial Vehicles (UAVs) are engaged in search-and-rescue flights in close proximity to humans, or when robots operate on a factory floor, precautionary measures need to be taken to ensure that the CPS adheres to its specified requirements and operates within its predefined safety envelope ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5a322-ec86-4fb2-b674-ed9494c1278d
• monitoring diverse properties at runtime ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5a327-e314-4079-af26-4d302f7c526a
• eterogeneity of hardware and software components within a CPS means that creating and implementing monitors, and subsequently collecting, processing, and checking the required data is often an arduous and time-consuming task. ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5a32f-9a3c-47e1-9e52-4aa47a591e4f
• source code instrumentation ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5e945-127f-4b55-9d7a-3cce21698738
• off-the-shelf monitoring approache ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5e968-b2fb-49e8-8e52-e8580b947151
• often provide inadequate support for instrumenting custom systems ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5e96e-d017-4521-b952-57c85266e0b3
• Few approaches have addressed the challenge of automatically generating customized, system-specific, runtime monitoring solutions and their maintenance and evolution support. ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b5e998-f6f6-4e35-88ae-01e73e73dd23
• When a new feature or functionality is introduced, the model is updated, and its respective code regenerated. As a result, MDE has been shown to improve productivity by enabling developers to specify a system at a higher level of abstraction ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b6dd9d-7eec-4314-9823-1fcf1c4f484a
• However, the application of MDE has fallen short of enabling the generation and evolution of a complete monitoring platform [16]. ls-type:: annotation hl-page:: 3 hl-color:: purple id:: 63b6dda5-192a-4345-9070-7ee6d73084f9
• Generating CPS Runtime Monitors ls-type:: annotation hl-page:: 3 hl-color:: green id:: 63b6ddd2-46eb-41c4-a61d-3387b2c0d3f4
• readily extended and updated when changes to the monitored system occur ls-type:: annotation hl-page:: 3 hl-color:: purple id:: 63b6dddf-5ecb-45d7-b5b8-393724736bf6
• parts of the system that are relevant for the monitoring infrastructure ls-type:: annotation hl-page:: 4 hl-color:: purple id:: 63b6de2a-ca80-4a46-a6bc-ab999d7b0aca
• This paper builds upon our earlier work [17, 18], in which we collected requirements for model-driven monitoring and derived an initial architecture. ls-type:: annotation hl-page:: 4 hl-color:: yellow id:: 63b6de38-c645-47c6-bc9a-90e54cc15ecb
• e evaluate our approach against a system for deploying UAVs for search-and-rescue and a second system that uses a set of mobile robots to assess indoor air quality. ls-type:: annotation hl-page:: 4 hl-color:: green id:: 63b6dea8-1a08-44f5-8911-b86abc792f77
• irst, we present a monitoring metamodel that enables us to generate a custom monitoring platform with a runtime model, instead of implementing it manually ls-type:: annotation hl-page:: 4 hl-color:: purple id:: 63b6deb4-c1eb-4a3e-bc87-57c1348dcb07
• econd, probes are automatically generated for the target technology ls-type:: annotation hl-page:: 4 hl-color:: purple id:: 63b6debe-f03c-4254-a86b-06f38ac4d5eb
• e can leverage different types of (off-the-shelf) constraint engines depending on the monitoring needs, ls-type:: annotation hl-page:: 4 hl-color:: green id:: 63b6df15-7ea3-43af-aebb-795cf2f86360
• nce a probe generator for a specific technology has been implemented, evolving the monitoring framework alongside the SuM is greatly simplified and only requires a few steps to update the model and automatically regenerate the platform1. ls-type:: annotation hl-page:: 4 hl-color:: green id:: 63b6df23-1605-4d30-8f03-71afeea590a0
• existing approaches are typically customized for a specific type of system, application domain, or support particular types of constraints ls-type:: annotation hl-page:: 5 hl-color:: green id:: 63b84bca-3d81-462d-8bbe-66989d750b4c hl-stamp:: 1673022412730
• significant upfront investment is required to implement and maintain custom monitoring infrastructures, often without inbuilt support for maintenance and evolution after deploymen ls-type:: annotation hl-page:: 5 hl-color:: purple id:: 63b84c07-ed8b-4501-bd4e-384ff614c0da hl-stamp:: 1673022476737
• Dronology, an open-source, multi-UAV system [31] with support for planning and executing UAV missions ls-type:: annotation hl-page:: 5 hl-color:: green id:: 63b84c53-bb7f-484f-88c2-537eaf2e33b2 hl-stamp:: 1673022549702
• constraint engine needs to be configured to check constraints, such as whether the UAVs remain within their designated altitude band, avoid no-fly zones, and are flight-worthy ls-type:: annotation hl-page:: 5 hl-color:: purple id:: 63b84c87-e7ab-40aa-967c-266d2f0124f6 hl-stamp:: 1673022602996
• if a new water sampling capability were introduced to Dronology, it would require additions and modifications to the hardware and software components ls-type:: annotation hl-page:: 5 hl-color:: purple id:: 63b84cf3-71d7-4f10-ad68-ab7b8300fad8
• New downward-facing sensors would be required to ensure that the UAV maintains a stable distance from the water during the collection process. ls-type:: annotation hl-page:: 5 hl-color:: purple id:: 63b84d01-b9bc-45ea-9aa3-4515958337f9
• to reduce the cost and effort of creating and maintaining a runtime monitor as illustrated for Dronology, it is necessary to (i) provide automated support for creating a system-specific solution, including the setup and configuration of the monitoring system and specification of its respective constraints, and (ii) avoid the additional effort of maintaining the monitoring system itself by enabling the infrastructure to co-evolve alongside the SuM. ls-type:: annotation hl-page:: 5 hl-color:: purple id:: 63b84d29-556c-4ed6-b174-8d1c904de7b5
• support ls-type:: annotation hl-page:: 6 hl-color:: purple id:: 63b84d3c-6ba7-4474-b84f-b57b7b8d25dd
• runtime monitoring architecture ls-type:: annotation hl-page:: 6 hl-color:: purple id:: 63b84d40-9e02-4247-87ae-b3ef322b30c7
• ease the task of maintaining and co-evolving monitors ls-type:: annotation hl-page:: 6 hl-color:: purple id:: 63b84d44-5f07-4517-a9ac-8f92143e2f6a
• one of the novel characteristics of GRuM is that for a SuM, both a Set of Probes, as well as a fully customized Monitoring Platform can be generated based on a model describing the parts to be monitored ls-type:: annotation hl-page:: 6 hl-color:: purple id:: 63b84d5f-e656-4b1d-ac2b-e1594fb26ac1
• MDE has been used extensively to generate applications and system components across a wide variety of domains, including automotive, railroad systems, business process engineering, and embedded system ls-type:: annotation hl-page:: 7 hl-color:: green id:: 63b8504a-bc50-42b1-a9e5-4bd230b6cb07 hl-stamp:: 1673023566926
• It only requires the parts of the SuM that need monitoring to be modeled, instead of the entire SuM ls-type:: annotation hl-page:: 7 hl-color:: purple id:: 63b8521a-5f39-43f0-b96f-b70cac48551a
• GRuM Monitoring Setup specifies exactly what parts of the system should be monitored and provides the mechanisms for performing the monitoring. ls-type:: annotation hl-page:: 7 hl-color:: green id:: 63b852a7-7364-4891-8f12-2dcb7a9de705
• monitors are typically defined and developed individually for each type of system or technology, such as byte-code instrumentation or dedicated service busses in service-based systems ls-type:: annotation hl-page:: 7 hl-color:: purple id:: 63b852c6-cddb-49b5-bf32-bcc6037f341c
• GRuM leverages MDE to describe the components of the SuM that need to be monitored, and then subsequently generates the monitoring infrastructure. To support this process we have created a dedicated MMM that can be used to define the relevant parts of the system for the resulting probes and the Monitoring Platform. ls-type:: annotation hl-page:: 7 hl-color:: purple id:: 63b852eb-9e43-4bf6-bfa9-68d88785b9ff
• [:span] ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63b85327-1a50-4bce-beae-3d3166e02e8e hl-type:: area hl-stamp:: 1673024292567
• monitoring configuratio ls-type:: annotation hl-page:: 7 hl-color:: green id:: 63b85343-2435-47bd-87ef-301dfe08d591
• ach agent has its own set of monitorable properties (MoProperty) ls-type:: annotation hl-page:: 7 hl-color:: green id:: 63b853d1-3270-4151-af33-442d8dc3f2aa
• Weaved Monitoring Model ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63b8544c-d455-4466-bbf7-dfcbc74bc71c
• determine ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63b8545c-7686-4e60-8ef5-b3e7585ea5b1
• Fragment ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63b85462-8942-47e0-a124-c2b18e70dba7
• Domain Model Fragment. Model weaving has been used in the past to connect di ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63b85463-7db9-4ebc-992d-cb38e3e81195
• Domain ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63b8546a-5a46-4bfd-96cb-585c2e5cb51d
• Weaved Monitoring Model (WM) ls-type:: annotation hl-page:: 8 hl-color:: green id:: 63bc209d-5ea9-44aa-8d25-7a9ab6c963c4
• MonitorableAgent ls-type:: annotation hl-page:: 9 hl-color:: green id:: 63bc20c2-f803-4aa9-817f-6b7efdf1124c
• MonitorableProperty ls-type:: annotation hl-page:: 9 hl-color:: green id:: 63bc20da-6333-4e6e-8105-ee54459adbd3
• a set of probes for collecting information from the system ls-type:: annotation hl-page:: 9 hl-color:: green id:: 63bc2108-6456-410f-a2d1-c1be07754983
• an instance of the Monitoring Platform for the SuM ls-type:: annotation hl-page:: 9 hl-color:: green id:: 63bc210e-3e69-402f-82a3-567964e08b39
• Listing 1: Excerpt of the WM for the Dronology system ls-type:: annotation hl-page:: 10 hl-color:: red id:: 63bc2121-6652-4fe7-b050-b8cdad6ab34d hl-stamp:: 1673273639057
• technology and language-specific set of probes ls-type:: annotation hl-page:: 10 hl-color:: green id:: 63bc23b3-6bcd-4642-b9a2-6f19513b316f
• can be reused for other Java-based systems ls-type:: annotation hl-page:: 10 hl-color:: yellow id:: 63bc23da-0963-4981-ae5c-53430a12cd50
• Model Query Engine ls-type:: annotation hl-page:: 10 hl-color:: green id:: 63bc2806-ba39-469d-bb8f-3f38e407f847
• monitoring platform ls-type:: annotation hl-page:: 11 hl-color:: green id:: 63bc2832-d384-4ab8-ab4a-532998064a8d
• Domain Model Fragment ls-type:: annotation hl-page:: 11 hl-color:: green id:: 63bc28e1-fadf-48da-86b8-ca5657cc5886
• monitoring concepts ls-type:: annotation hl-page:: 11 hl-color:: green id:: 63bc28ec-84f8-492c-a609-61078e074c0a
• . Platform and Probe Generation ls-type:: annotation hl-page:: 12 hl-color:: green id:: 63bc2905-d93e-4942-8736-3ac187f3af8d
• Platform Usage & Runtime Monitoring ls-type:: annotation hl-page:: 12 hl-color:: green id:: 63bc290b-1633-4f88-b90b-fea5bdea3f31
• integrated directly by the developers, or as part of the continuous integration process ls-type:: annotation hl-page:: 12 hl-color:: green id:: 63bc295d-550f-4de2-b1cb-73509e58fa22
• MoProperties ls-type:: annotation hl-page:: 12 hl-color:: green id:: 63bc2975-3d85-4834-88ce-664f9e25dad5
• middleware ls-type:: annotation hl-page:: 12 hl-color:: red id:: 63bc29c7-2d78-45e7-b64b-6960153edc1b hl-stamp:: 1673275851592
• System Evolution ls-type:: annotation hl-page:: 13 hl-color:: green id:: 63bc2c03-c483-48ce-9129-94f69843e64d
• easily incorporated into the monitoring platform as well ls-type:: annotation hl-page:: 13 hl-color:: yellow id:: 63bc2c1a-34b8-4d0b-be34-e1cbb7837ab6
• if an additional property of an existing agent needs to be added to the set of monitored properties, this change can be directly performed at the model level ls-type:: annotation hl-page:: 13 hl-color:: green id:: 63bc2c88-d71f-4f10-b24a-a1f90473d7f0 hl-stamp:: 1673276556244
• probes and platform code can be completely regenerated without the need to manually adapt the code. ls-type:: annotation hl-page:: 13 hl-color:: green id:: 63bc2c96-14c6-4b29-b36d-ecb07f300a72
• we used XText [48] to define a textual domain-specific language (DSL) for the WM (see example in Listing 1) which allows easy creation and modification of the WM for a specific SuM. ls-type:: annotation hl-page:: 13 hl-color:: green id:: 63bc2cc9-913c-4875-ba32-16a44e4ca432
• two distinct code generators to support two different target languages and technologies ls-type:: annotation hl-page:: 13 hl-color:: green id:: 63bc2cf1-9729-4c1a-b1f4-935e9a74f5b9
• At runtime, an instance of the Domain Model Fragment is created and used as the runtime mod ls-type:: annotation hl-page:: 14 hl-color:: green id:: 63bc34c6-ecf6-4257-949c-8cc6368eedbd
• while Viatra provides tools and incremental evaluation [50, 51], it does not support more complex constraints, such as temporal ones ls-type:: annotation hl-page:: 14 hl-color:: green id:: 63bc34ef-60cb-4d02-9f1d-5c650e421622
• Complex Event Processing (CEP) ls-type:: annotation hl-page:: 14 hl-color:: green id:: 63bc34f6-af00-4441-b614-31c647f4bfae
• reasonable effort ls-type:: annotation hl-page:: 15 hl-color:: green id:: 63bc350a-20f4-4023-b84c-a21d3d59637f
• ssessing the effort ls-type:: annotation hl-page:: 15 hl-color:: green id:: 63bc3520-d8e4-428d-94ad-1d061aca2075
• create a runtime monitoring platform ls-type:: annotation hl-page:: 15 hl-color:: green id:: 63bc353c-0697-49fe-b5b0-7201729e3258
• efficiently monitor timesensitive runtime data ls-type:: annotation hl-page:: 15 hl-color:: green id:: 63bc3554-ca84-4bd0-aafa-bccd035bf95d
• measuring the model-set-latency of the monitoring platform ls-type:: annotation hl-page:: 15 hl-color:: green id:: 63bc355c-1fe9-4c95-b898-c3da06873b95
• [:span] ls-type:: annotation hl-page:: 15 hl-color:: green id:: 63bc35c8-e461-49f2-bb21-0894eb707c50 hl-type:: area hl-stamp:: 1673278967744
• [:span] ls-type:: annotation hl-page:: 15 hl-color:: blue id:: 63bc35db-b13e-4b50-8786-1b1247c510ae hl-type:: area hl-stamp:: 1673278963252
• One was assigned to Dronology and one to the TurtleBot system. We then recorded the time taken to complete each activity ls-type:: annotation hl-page:: 16 hl-color:: green id:: 63bc3625-fdfd-41bb-8927-1eb4ae8a2a23
• “thinking-time” which is expected to be similar whether the monitor is built from scratch or using GRuM ls-type:: annotation hl-page:: 16 hl-color:: green id:: 63bc36b3-4eda-4775-b6be-0c8f7c49ad99
• Case 1 – Dronology ls-type:: annotation hl-page:: 16 hl-color:: yellow id:: 63bc36d5-0483-4755-a056-2791ccff2036
• . Altogether, the implementation of constraints and extensions was completed within about 20 hours ls-type:: annotation hl-page:: 16 hl-color:: green id:: 63bc373d-6509-4c88-b90f-92461a5feab0
• Case 2 – ROS TurtleBot ls-type:: annotation hl-page:: 17 hl-color:: yellow id:: 63bc8375-d23b-4bbc-9243-2d09308bbaf7
• MQTT-Forwarder for sending data to the monitoring platform ls-type:: annotation hl-page:: 17 hl-color:: green id:: 63bc8640-7bfa-4af3-907b-46e1ac2035c9
• we developed this application from scratch without prior experience with the system or technolog ls-type:: annotation hl-page:: 17 hl-color:: green id:: 63bc8656-7c76-4036-95f7-b7ce1c5b9db4
• Domain Model Fragment (1) for the TurtleBot consisted of 10 properties with 23 attributes ls-type:: annotation hl-page:: 17 hl-color:: green id:: 63bc8664-4d45-4a74-8799-cb3b5ffcda16
• 5 hours were spent implementing the GRuM monitor after properties had been selected. ls-type:: annotation hl-page:: 17 hl-color:: green id:: 63bc8676-efb4-451c-9da0-1b829fa35207
• 7 constraints ls-type:: annotation hl-page:: 17 hl-color:: yellow id:: 63bc8686-dc99-41a6-82de-3a2ce739f55c hl-stamp:: 1673299592488
• add monitorable properties to a system, perform constraint checks using the newly collected data, update an existing element (e.g., change the information and/or type of a property) and remove an existing element from the model. ls-type:: annotation hl-page:: 18 hl-color:: green id:: 63bc86cb-beb5-48e0-994c-f4f7cf397cff
• (i) a new property ls-type:: annotation hl-page:: 18 hl-color:: green id:: 63bc871e-d222-4e37-b7ad-e572d117e3ce
• (ii) an existing property is modified ls-type:: annotation hl-page:: 18 hl-color:: green id:: 63bc8722-c6fd-439a-a364-7f2ed534e90e
• (iii) a property is removed ls-type:: annotation hl-page:: 19 hl-color:: green id:: 63bc8729-ad90-49ff-85c2-13e801b16ecd
• GRuM provides a high degree of automation, reducing the manual effort that is required to adopt changes in the SuM in the monitoring platform ls-type:: annotation hl-page:: 20 hl-color:: green id:: 63bc8844-2328-4e0e-a762-242a28dabef1
• UAV search mission scenario ls-type:: annotation hl-page:: 20 hl-color:: green id:: 63bc88ce-0356-485a-b835-9fb0758039e8
• multiple UAVs search for a missing person in a predefined are ls-type:: annotation hl-page:: 20 hl-color:: green id:: 63bc88d3-7c15-48e1-b7a2-577a904f39dd
• he monitoring infrastructure was set up on a standard Desktop Computer, with 16GB of RAM, running Ubuntu 20, using Java 11, and Ecore2.23. We randomly assigned 5 flight routes to each UAV, and then collected runtime data and performed constraint checks ls-type:: annotation hl-page:: 20 hl-color:: green id:: 63bc8a4b-25b3-4de5-8282-ec5ba0a98ec1
• violations were reported correctly ls-type:: annotation hl-page:: 20 hl-color:: yellow id:: 63bc8a59-5a7e-42df-9bfb-8d4907972f2a hl-stamp:: 1673300571717
• 8 Speed Reduction: When the bot operates below 25% battery level, the speed must not exceed 2.0m/s anymore ls-type:: annotation hl-page:: 21 hl-color:: yellow id:: 63bc8ab5-04aa-41e8-874a-2daabadc35cb
• Measurements Accuracy: To ensure accurate measurements, an alert should be raised when measurements are transmitted while movin ls-type:: annotation hl-page:: 21 hl-color:: yellow id:: 63bc8abf-1f59-4501-b889-c24a41460417
• temporal constraints with Esper ls-type:: annotation hl-page:: 17 hl-color:: yellow id:: 63bd3b61-ad54-4cd7-ac3b-dc9e9e023335
• properties ls-type:: annotation hl-page:: 19 hl-color:: red id:: 63bd3dac-e5ae-49f4-85e9-8f01bca6086b hl-stamp:: 1673346492075
• While these were created by several authors of this paper ls-type:: annotation hl-page:: 25 hl-color:: red id:: 63bd3e0b-459b-451d-b8b6-9cf6a59f5ccb
• differences in terms of the technology used, their architectural styles, and propertie ls-type:: annotation hl-page:: 25 hl-color:: red id:: 63bd3e3b-ce26-459a-be0d-55e85b21618c
• While we have been using a simulator for the Dronology use case, Ardupilot is a high-fidelity simulator providing a realistic execution environment, and our previous experience has confirmed that this closely resembles real-world scenarios ls-type:: annotation hl-page:: 25 hl-color:: green id:: 63bd3eaf-efaf-4f43-8f73-1d18b7c7e705 hl-stamp:: 1673346742690
• (1) describe relevant properties ls-type:: annotation hl-page:: 26 hl-color:: green id:: 63bd3ee0-946a-4865-b841-e943c61cdb66
• (2) collect and check runtime data via the automatically generated monitoring platform ls-type:: annotation hl-page:: 26 hl-color:: green id:: 63bd3ee6-1d8a-4d48-99ca-d62b084570eb
• automatically generating runtime monitors ls-type:: annotation hl-page:: 28 hl-color:: green id:: 63bd3f87-6ff3-4b79-b468-02c18c85df1b
• SuM’s technology and architectural style ls-type:: annotation hl-page:: 28 hl-color:: yellow id:: 63bd3f97-8395-4b79-b765-50f7e9ffe7fb