Notesale: Turn your study into money

Document Preview

Extracts from the notes are below, to see the PDF you'll receive please use the links above

---

PETER MUNIU KURIA
MAC-1-1664-1/2011
DISTRIBUTED SYSTEMS
CISY 432
1
...
Example:Active Directory data store as a single, general-purpose database for all of an
organization’s data
...


Discuss the role of a server replica manager in a

distributed system?
Replication Manager facilitates the administration of replica sets over LANs,
WANs, and the Internet
...
IT pros can readily manage
distributed replicas on laptops through Replication Manager's indirect
synchronization capabilities
...
The replica manager also communicates with others replica managers
running on the other replica servers
...


Explain the two types of inconsistencies that may

occur in a data store



It usually occurs because of incomplete database transactions
...
Corruption could
be the outcome of virus infection, improper system shutdown,
application malfunction, and other similar reasons
...
e
...




operations are performed in different orders at different replicas
causing even grave problems than staleness because applying
operations in different orders can lead to different results that
cannot be consolidated
...


What is a conflicting operation?
Conflicting operation is a situation that occurs when clients on different
machines access data store concurrently
...
g Concurrent read-write,
write-read and write-write operations by two transactions are conflicting
...


Differentiate between data centric and client

centric consistence models
Data-centric is a class in consistency model; these are consistency models
that apply to the whole data store
...
The strict
consistency model requires that any read on data item returns a value
corresponding to the most recent write on the data item
...


6
...

(a) Strict Consistency
"The strict consistency model is the strongest form of memory coherence,
having the most stringent consistency requirements
...
A problem with strict consistency is that the interpretation of most recent
isn’t clear in a distributed data store
...
Due to the fact that it is not

possible to achieve absolute global time in a distributed system and the fact
that communication between replicas can never be instantaneous, strict
consistency is impossible to implement in a distributed data store
...

In a data store that provides sequential consistency, all clients see all (write)
operations performed in the same order
...
The only requirement is that
all clients see the same total ordering
...
The exact order in which
the memory access operations are interleaved does not matter
...
three operations and another process sees a
different one, the memory is not a sequentially consistent memory
...
relaxes the requirement of the sequential
model for better concurrency
...
Two
writes are casually related if the execution of one write possibly influences the
value written by the second write
...
A write of a particular
data item is followed by a read of that data item in any client
...
Concurrent writes can be
executed in any order, as long as program order is respected
...
Before synchronization access can be
performed, all previous regular data accesses must be completed
...
This essentially leaves the problem of consistency up to the
programmer
...
" This model defines
...

(e) Release Consistency Model:
"Release consistency is essentially the same as weak consistency, but
synchronization accesses must only be processor consistent with respect to
each other
...
All pending acquires (e
...
, a lock operation)
must be done before a release (e
...
, an unlock operation) is done
...

"Release consistency is a further relaxation of weak consistency without a
Significant loss of coherence
...
variants of release consistency, it requires the programmer (or
compiler) to use acquire and release at the start and end of each critical
section, respectively
...
If it is desired that
elements of an array be accessed independently in parallel, then different array
elements must be associated with different locks
...
"

7
...


8
...
Also
show whether the protocol is a local-write or a
remote write (with justifications)
...
All write operations on a data item
are forwarded to the server holding that item’s primary copy
...

Advantages
This protocol is easy to implement
Desadvantages
It does not scale well and has a negative impact on performance
...


Primary-Backup
Local –write
– Allow local reads, sent writes to primary
– Block on write until all replicas are notified
– Implements sequential consistency
--writes can be performed at a data item’s primary copy
The primary-backup protocol allows reads to be executed at any replica;
however, writes can still be performed at a data item’s primary copy
...

Advantages
Because of the blocking write, this protocol can easily be used to
implement sequential consistency
...

Desadvantages
This protocol has negative impact on performance and scalability
...

Migration
Local -write
The migration protocol is the first of the local-write protocols
...
However, when a
data item is accessed it is moved from its original location to the replica of
the client accessing it
...

Disadvantages
Concurrent reads and writes can lead to thrashing behavior where the
data item is constantly being copied back and forth
...

This improves on the write performance of primary-backup, and avoids
some of the thrashing of the migration approach
...
Before
disconnecting from the network the client becomes the primary allowing it
to perform updates locally
...

Active Replication
The active replication protocol is a replicated write protocol
...
The writes can be propagated using either point-to-point
communication or multicast
...

Disadvantages
Requires an atomic multicast or a centralized sequencer, neither of which
are scalable approaches

---