Week 1

Q1. A supply system and inventory system. A system where employees place order for parts and order alert prompts an employee in inventory to draw the item and send it to movement. The inventory system updates its holdings (and if holdings are low will place an order with the wholesaler). The physical part is moved to movement the system prompts the movers to ship the item to the client. The system allows the client to track the order through the system. The mover delivers the order and closes the ticket on the system. The client confirms receipt of the order. In this example all the users do not track the different systems that are involved with allowing this connectivity.

Q2. Middleware in a distributed system is software that is layered below the application layer and above the operating system layer. It allows various applications to communicate with each other it also allows applications to run on different operating systems.

Q3. Still working on it

Q4. To the user the fact that the system is distributed or not is unknown. The user isn’t aware that some of the processes s/he uses could be located a great distance away from them. If a system is able to appear to both users and applications to be one computer then the distributed system is transparent. There are 6 types of transparency: access, location, migrating, relocation, replication, concurrency, and failure. An example of failure transparency is when an application or client within the system fails but the other clients never know that there is an issue.

Q5. A big reason failure ad recovery transparency is difficult is there is no way to distinguish between a dead component and a really slow component. The example provided in the reading talks about a web server which might be processing millions of requests and is running slow. The user’s browser will timeout and report an error that the server isn’t responding. From this it’s not clear if the web server is offline or just really slow to reply.

Q6. The times where not being completely transparent is useful is when a service is located very far from the user/client. In line with examples from the reading a military example is, when we deploy networks around the world, we cannot always deploy servers and therefore a mission in Egypt only supports a few local services but the bulk of SharePoint applications and others are hosted in Canada. Therefore, the speed of the requests between the mission in Africa and Canada is slower ten if the services were provided locally. Users can expect that their connection will be slow and won’t be so angry when they have to wait a second or two to access a file.

Q7. An open distributed system is a system that offers services according to standard rules that describe the syntax and semantics of those services. Benefits that an open system provides is different develops can design products that create different systems but the different systems are interoperable because they all follow the set standard rules for syntax etc.

Q8. A scalable system is that can increase or decrease according to the requirement. The system will function in the same manner and the users do not notice if the system grows to add more users, expands to locations around the world or the amount of administrators decreases. An example is the DWAN for the military. The DWAN core servers are in Canada but there are remote servers and clients all over the world. As the military moves around DWAN nodes connect and disconnect all the time, increase the number of clients and decrease the number clients.

Q9. The techniques outlined in the reading for scalability are hide communication latencies, distribution and replication.

Q10. Within grid computing systems virtual organizations group users together to support collaboration between the different groups or businesses. To implement virtual organizations users and applications are provided rights to access data and resources. To me this would eb something like SharePoint where different units are provided rights to access SharePoint pages where they can share products/reports with other units who are also part of the virtual organization.

Q11. Transaction abortion doesn’t reset the world when the transaction results in a physical output. If a transaction resulted in something getting printed, and the printer completed the print job before the abort signal was sent then that cannot be undone or reset. 

Q12. A coordinator will ensure that the nested transactions will commit when they can and if a nested transaction aborts then the other nested transactions also abort.

Q13. We argued that distribution transparency may not be in place for pervasive systems. This statement is not true for all types of transparencies. Give an example. An example of when transparency should be in place for pervasive systems is when moving from one group to another such as a mobile phone which crosses from one access point to another there is a handover action that is transparent to the user.

Q14. Extend the list with more examples of distributed pervasive systems: mobile networks; cars/trucks/vehicles; GPS;

 Explain the following terms:
Grid computing - An example is the NATO FMN (Federated Mission Network). Grid computing is where distributed systems which are under different administration, have various hardware and software. The various systems are connected to allow for collaboration. The FMN is very much like this, where different nations have their own systems that they connect to one another in a tactical environment so partnering nations can communicate and conduct operations together.
Virtual organization – within grid computing the virtual organization is the group of users/clients who have rights to access and share data within the collaboration spaces created by the grid computing.
ACID – Refers to the characteristic properties of a transaction. Each letter refers to a characteristic. A: atomic; C: consistent; I: isolated; and D: durable.
EAI – Enterprise Application integration. A topic which focuses on developing solutions to allow different applications to work together or communicate with each other.
Pervasive computing – Another term could be mobile computing. The distributed system that deals with pervasive computing is focused on mobile devices from wearable tech to cellphones, cars etc.


Week 2:
4 styles have been identified for distributed systems, they are: Layered architecture; object-based architecture; data-centered architecture; and event-based architecture.

1. One way is to replicate the service of the origin server to an edge server which is closer to the client. This will reduce the perceived latency by the client. Although on the backend where the user doesn’t see this can cause other issues within distributed systems in maintaining the most up to date information on all servers who hold replica information.

2. A three tired architecture is where the processes of a distributed systems are hosted on different platforms. An example would be to have the user interface on the client platform, the processing of data on a data server, and perhaps storage on a third platform.

3. Vertical distribution is where the multitiered architecture is split by logical layers on different machines. Horizontal distribution is taking one layer and splitting it across multiple machines.

4. A main problem is P1 has to wait for P2 to Pn to process the request. If n is large then it will take a long time for a reply to be provided to P1. Also if there is an error or one of the processes is not function/not there, you may never see a reply as Pi-1 is always dependant on Pi.

5. From the reading it sees that a disadvantage is when a node wants to leave the structured overlay network. Repartitioning the space is difficult and symmetry is can be lost.

15. Examples of feedback control systems in cars are: speed – the user provides input by pressing the gas petal and the system through feedback adjusts the air to gas ratio in the engine to make the engine work harder and increase the speed); Cruise control – is a better example as the car has to keep checking it’s speed (the driver doesn’t do the checking) and either speed up or slow down in order to maintain the set speed; Engine temperature – the engine is to be at a set temperature so the system monitors the engine temperature. If the engine is too hot then coolant is added; Changing gears (automatic) – the engine compares the current rpms to a standard. As the engine speeds up or slows down the rpms change. Depending on the change the engine will shift gears.

16. An example of a self-managing system where the analysis component is completely distributed is a mobile phone. The phone when is passes out of range of one tower and into range of another town will automatically conduct a handshake between the different towers to maintain a connection. This is done without the user knowing.