DISTRIBUTED SYSTEMS MIDTERM EXAM QUESTIONS AND
ANSWERS
Openness - ANSWER Compatibility with other resources
Interoperability - ANSWER Open distributed systems offer components that
can be used by or integrated into other systems
Portability - ANSWER System should be flexible enough for apps to be
compliant with other distributed systems
Scalability (size) - ANSWER More users and resources may be added without
noticeable loss of performance
Scalability (geographically) - ANSWER Despite the distance between users
and resources communication delays are hardly noticed
Scalability (administrative) - ANSWER Easily manageable even if spanning
many independent administrative organisations
3 types of scaling out techniques - ANSWER 1. Hiding communication
latencies
2. Partitioning and Distribution
3. Replication
Hiding Communication Latencies - ANSWER * The goal is to reduce the wait
time for the remote service requests of clients
* Requires asynchronous communication
* Fails for interactive services where each step requires approval from the
server
** Work around is to outsource the job of making checks to the client itself and
limiting the communication required by the server to a minimum
Partitioning and distributing - ANSWER Splitting a component into smaller
parts and spreading out across the system
Examples of partitioning and distributing - ANSWER * Internet Domain
Name System
, * Document handling over the internet
** The web itself is physically partitioned and distributed across many servers,
each handling a number of docs
** The name of the server that handles a document is encoded into the
document's URL
Replication - ANSWER Increases availability and helps to balance the load
between components leading to better performance, helps with hiding
communication latency by providing copies(replicas)
* Drawback is that resources can be inconsistent, meaning that the browser
returns a cached document of which validity has not been checked in the last
few minutes
Bad for stock exchanges, ebay, etc.
Caching - ANSWER When replication is done near the client
3 types of high performance computing (HPC) - ANSWER 1. Cluster
computing
2. Grid computing
3. Cloud computing
3 types of pervasive systems - ANSWER 1. Ubiquitous computing
2. Mobile computing
3. Sensor networks
Cluster computing - ANSWER Type of HPC- single program is run in parallel
on multiple machines. Each cluster consists of nodes with 1 single master node.
Ex. linux beowulf
Grid computing - ANSWER Isolated group of computers with a hierarchy of
access and resources. Fabric layer, provides interface
4 layers of grid computing - ANSWER 1. Fabric layer- provides interfaces to
local resources
2. Connectivity layer- consists of communication protocols for supporting grid
transactions that span the usage of multiple resources
3. Resource layer- manages a single resource
ANSWERS
Openness - ANSWER Compatibility with other resources
Interoperability - ANSWER Open distributed systems offer components that
can be used by or integrated into other systems
Portability - ANSWER System should be flexible enough for apps to be
compliant with other distributed systems
Scalability (size) - ANSWER More users and resources may be added without
noticeable loss of performance
Scalability (geographically) - ANSWER Despite the distance between users
and resources communication delays are hardly noticed
Scalability (administrative) - ANSWER Easily manageable even if spanning
many independent administrative organisations
3 types of scaling out techniques - ANSWER 1. Hiding communication
latencies
2. Partitioning and Distribution
3. Replication
Hiding Communication Latencies - ANSWER * The goal is to reduce the wait
time for the remote service requests of clients
* Requires asynchronous communication
* Fails for interactive services where each step requires approval from the
server
** Work around is to outsource the job of making checks to the client itself and
limiting the communication required by the server to a minimum
Partitioning and distributing - ANSWER Splitting a component into smaller
parts and spreading out across the system
Examples of partitioning and distributing - ANSWER * Internet Domain
Name System
, * Document handling over the internet
** The web itself is physically partitioned and distributed across many servers,
each handling a number of docs
** The name of the server that handles a document is encoded into the
document's URL
Replication - ANSWER Increases availability and helps to balance the load
between components leading to better performance, helps with hiding
communication latency by providing copies(replicas)
* Drawback is that resources can be inconsistent, meaning that the browser
returns a cached document of which validity has not been checked in the last
few minutes
Bad for stock exchanges, ebay, etc.
Caching - ANSWER When replication is done near the client
3 types of high performance computing (HPC) - ANSWER 1. Cluster
computing
2. Grid computing
3. Cloud computing
3 types of pervasive systems - ANSWER 1. Ubiquitous computing
2. Mobile computing
3. Sensor networks
Cluster computing - ANSWER Type of HPC- single program is run in parallel
on multiple machines. Each cluster consists of nodes with 1 single master node.
Ex. linux beowulf
Grid computing - ANSWER Isolated group of computers with a hierarchy of
access and resources. Fabric layer, provides interface
4 layers of grid computing - ANSWER 1. Fabric layer- provides interfaces to
local resources
2. Connectivity layer- consists of communication protocols for supporting grid
transactions that span the usage of multiple resources
3. Resource layer- manages a single resource
