From the Edge to the Cloud - Research and Engineering challenges for IoT systems
Schahram Dustdar, TU Wien, Austria.
Smart Evolution - People, Services and Things
- in the past, people were outside of the system's development. Nowadays, Information Systems are actually a composition of these three elements: people, services and things.
E.g. Smart energy networks, Smart Government and administration, Smart housing systems etc. All these systems are interconnected
---
He makes an analogy with the autonomic nervous system, connecting different organs, and having messages being sent from one organ to the other. Moreover, whenever there is something wrong, some signal is sent so that the problem may be addressed.
- In ISs, sensors are being added slowly to the current information systems. Previously they were not there.
---
Think Ecosystems: people, systems and things
1. Robustness and Resilience mechanism: achieving stability in the presence of disruption
2. Measures of health: diversity, population trends, other key indicators
- Monitoring is not ubiquitous today. We must create mechanisms to verify if the ISs are healthy
----
Smart City are good examples of how the ISs landscape is leading to Everything as a Service (EaaS)
---
He mentioned an initiative called Society 5.0 from Japan, which aims at rethinking society and economy in this new reality we are facing.
They want to think about society evolution in terms of:
Hunting society - Agrarian Society - Industrial Society - Information Society - Super Smart Society
---
"The scientists of today think deeply instead of clearly. One must be sane to think clearly, but one can think deeply and be quiet insane" - Nikola Tesla
---
He challenges the general assumption that we have a linear history. E.g. ancient computers, such as the Stonehenge; and ancient pyramids have a quite similar structure of computer architecture.
These ancient societies were not as primitive as we think. We need to be aware of something that has to do with education, which may be exemplified by these two quotes:
"The purpose of education is to replace an empty mind with an open one" - Malcolm Forbes
"We cannot teach people anything, we can only help them discover it within themselves" - Galileo
---
Let's now examine the assumptions, models and abstractions that we want to create from now on.
We should provide consciousness and creativity support - Architecture of values
---
The kinds of ecosystems that we need to think about are composed of architecture, structure and dynamics. All three elements must be modeled.
Three new paradigms can help us build models for them:
1) Elastic Systems
Using a metaphor of physical elasticity to build systems, such that we have two states:
stretch - acquire new resources, reduce quality
shrink - release resources, increase quality
You can talk about a three dimensional space composed of the following dimensions: resource elasticity; quality elasticity; and cost and benefits. Work by Dustdar S. et al. (2011) Principles of Elatic Processes. IEEE Internet Computing, 15 (5).
He describes a Phd work in which they create a kind of a programing language having a taxonomic structure describing Elasticity directive primitives.
He also explained the following works
SYBL: an Extensible Language for Controlling Elasticity in Cloud Applications.
MELA: Monitoring and Analyzing Elasticity of Cloud Services.
(I did not write the whole reference)
2) Social Compute Units (SCUs)
How do you find the right people and put them together to deliver or to consume a particular service.
They are developing the concept of a Social Compute Unit (kind of CPU) which has primitives to operate on teams, such as: Create teams, Dissolve teams, Scale up teams, Scale down teams, Merge teams, Compose teams, Split teams and Cluster teams.
We should also look at Algorithmic Team Formation: there is today a variety of algorithms to create teams. They should be investigated closely, including the ethical dimensions to avoid problems that are common in our world, such as racism and exclusion.
3) Osmotic Computing
Dynamic management of (micro)services taking the perspective of the chemical osmosis phenomenon.
---
Perspectives on the IoT
Cloud-centric perspective
- Assumptions - cloud provides core services; edge provides local proxies for the cloud (offloading parts of the cloud's workload)
- This led us back to centralized services, simply because the edges are assumed to play a supportive role.
- But not every service should be provided by the cloud.
This led to a better perspective:
Internet-centric perspective
- Internet is the center of IoT architecture; edges devices are gateways to the Internet (not to the Cloud)
- Thus, things belong to partitioned subsystems and LANs rather than a centralized system.
---
Traditional ICT view vs. Societal View on Smart City: focus on the infrastructure vs. focus on the involvement of the community
Holistic view Architectures of values
- inclusion of all stakeholders
- integrated management of physical, ICT and social infrastructure
- generation of new value
He proposes the creation of Cyber-Human Smart City Values which include:
- societal values: direct inclusion and empowerment of citizens as key stakeholders of the city; and
- business values: new labor/work models supported by incentive mechanisms, team formation algorithms and negotiation protocols; and new business models.
He poses some research challenges that comprise social, ethical and technical challenges.
---
We should have a co-existence of AI and human beings, truly collaborative. It is a mistake to think that AI should dictate how things are done; or will solve everything.
We should seek cyber-coexistence - not only machines acting on our behalf, but we must create a partnership model for IoT/Edge/Cloud
He mentions some interesting references (we should definitely get the slides!)
--
Q&A:
He agrees with Oscar that we need conceptual model, but conceptual modeling should provide practical approaches, because we need to create these systems very fast, so there is not much time to spend on modeling alone. This is the dilemma. But we do need to go deep on conceptual modeling to reflect what we are becoming, and what we do. We should be reflecting on the problems existing in our society, the wrong viewpoints on how and why software is made (the paradigm of making and advertising software is a mistake).
We must be aware that "fear" leads us to go wrong on how we produce software. For instance, the fear that robots will replace people in jobs and such.
We are not powerless, we should be able to get ahead in the game, and not be led by the economic interests and other agents that are now driving the way things are done.
The tenured professors should step up and get active in social science publication means, and basically make people aware of what technology can do. The fragmentation of science in technical and social sciences is a complete nonsense. We should realize that there is no progress in science which is not multidisciplinary.
Smart Evolution - People, Services and Things
- in the past, people were outside of the system's development. Nowadays, Information Systems are actually a composition of these three elements: people, services and things.
E.g. Smart energy networks, Smart Government and administration, Smart housing systems etc. All these systems are interconnected
---
He makes an analogy with the autonomic nervous system, connecting different organs, and having messages being sent from one organ to the other. Moreover, whenever there is something wrong, some signal is sent so that the problem may be addressed.
- In ISs, sensors are being added slowly to the current information systems. Previously they were not there.
---
Think Ecosystems: people, systems and things
1. Robustness and Resilience mechanism: achieving stability in the presence of disruption
2. Measures of health: diversity, population trends, other key indicators
- Monitoring is not ubiquitous today. We must create mechanisms to verify if the ISs are healthy
----
Smart City are good examples of how the ISs landscape is leading to Everything as a Service (EaaS)
---
He mentioned an initiative called Society 5.0 from Japan, which aims at rethinking society and economy in this new reality we are facing.
They want to think about society evolution in terms of:
Hunting society - Agrarian Society - Industrial Society - Information Society - Super Smart Society
---
"The scientists of today think deeply instead of clearly. One must be sane to think clearly, but one can think deeply and be quiet insane" - Nikola Tesla
---
He challenges the general assumption that we have a linear history. E.g. ancient computers, such as the Stonehenge; and ancient pyramids have a quite similar structure of computer architecture.
These ancient societies were not as primitive as we think. We need to be aware of something that has to do with education, which may be exemplified by these two quotes:
"The purpose of education is to replace an empty mind with an open one" - Malcolm Forbes
"We cannot teach people anything, we can only help them discover it within themselves" - Galileo
---
Let's now examine the assumptions, models and abstractions that we want to create from now on.
We should provide consciousness and creativity support - Architecture of values
---
The kinds of ecosystems that we need to think about are composed of architecture, structure and dynamics. All three elements must be modeled.
Three new paradigms can help us build models for them:
1) Elastic Systems
Using a metaphor of physical elasticity to build systems, such that we have two states:
stretch - acquire new resources, reduce quality
shrink - release resources, increase quality
You can talk about a three dimensional space composed of the following dimensions: resource elasticity; quality elasticity; and cost and benefits. Work by Dustdar S. et al. (2011) Principles of Elatic Processes. IEEE Internet Computing, 15 (5).
He describes a Phd work in which they create a kind of a programing language having a taxonomic structure describing Elasticity directive primitives.
He also explained the following works
SYBL: an Extensible Language for Controlling Elasticity in Cloud Applications.
MELA: Monitoring and Analyzing Elasticity of Cloud Services.
(I did not write the whole reference)
2) Social Compute Units (SCUs)
How do you find the right people and put them together to deliver or to consume a particular service.
They are developing the concept of a Social Compute Unit (kind of CPU) which has primitives to operate on teams, such as: Create teams, Dissolve teams, Scale up teams, Scale down teams, Merge teams, Compose teams, Split teams and Cluster teams.
We should also look at Algorithmic Team Formation: there is today a variety of algorithms to create teams. They should be investigated closely, including the ethical dimensions to avoid problems that are common in our world, such as racism and exclusion.
3) Osmotic Computing
Dynamic management of (micro)services taking the perspective of the chemical osmosis phenomenon.
---
Perspectives on the IoT
Cloud-centric perspective
- Assumptions - cloud provides core services; edge provides local proxies for the cloud (offloading parts of the cloud's workload)
- This led us back to centralized services, simply because the edges are assumed to play a supportive role.
- But not every service should be provided by the cloud.
This led to a better perspective:
Internet-centric perspective
- Internet is the center of IoT architecture; edges devices are gateways to the Internet (not to the Cloud)
- Thus, things belong to partitioned subsystems and LANs rather than a centralized system.
---
Traditional ICT view vs. Societal View on Smart City: focus on the infrastructure vs. focus on the involvement of the community
Holistic view Architectures of values
- inclusion of all stakeholders
- integrated management of physical, ICT and social infrastructure
- generation of new value
He proposes the creation of Cyber-Human Smart City Values which include:
- societal values: direct inclusion and empowerment of citizens as key stakeholders of the city; and
- business values: new labor/work models supported by incentive mechanisms, team formation algorithms and negotiation protocols; and new business models.
He poses some research challenges that comprise social, ethical and technical challenges.
---
We should have a co-existence of AI and human beings, truly collaborative. It is a mistake to think that AI should dictate how things are done; or will solve everything.
We should seek cyber-coexistence - not only machines acting on our behalf, but we must create a partnership model for IoT/Edge/Cloud
He mentions some interesting references (we should definitely get the slides!)
--
Q&A:
He agrees with Oscar that we need conceptual model, but conceptual modeling should provide practical approaches, because we need to create these systems very fast, so there is not much time to spend on modeling alone. This is the dilemma. But we do need to go deep on conceptual modeling to reflect what we are becoming, and what we do. We should be reflecting on the problems existing in our society, the wrong viewpoints on how and why software is made (the paradigm of making and advertising software is a mistake).
We must be aware that "fear" leads us to go wrong on how we produce software. For instance, the fear that robots will replace people in jobs and such.
We are not powerless, we should be able to get ahead in the game, and not be led by the economic interests and other agents that are now driving the way things are done.
The tenured professors should step up and get active in social science publication means, and basically make people aware of what technology can do. The fragmentation of science in technical and social sciences is a complete nonsense. We should realize that there is no progress in science which is not multidisciplinary.
Nenhum comentário:
Postar um comentário