report on recovery in mobile database system

RECOVERY IN MOBILE DATABASE SYSTEM

TABLE OF CONTENT

ABSTRACT

INTRODUCTION

PREVIOUS WORK

SYSTEM MODEL

PROBLEM SPECIFICATION

METHODOLOGY

PROPOSED RECOVERY SCHEME

MERITS AND DEMERITS

FUTURE ASPECTS

CONCLUSION

REFERENCES

ISE Page 1 DSCE


ABSTRACT

Mobile computing is a rapidly emerging trend in distributed computing. This
new mobile computing environment poses challenging problems in designing
fault-tolerant systems because of the dynamics of mobility, and limited
bandwidth available on wireless links. Traditional fault-tolerance schemes,
therefore, cannot be directly applied to these systems. Mobile systems are often
subject to environmental conditions which can cause loss of communication or
data. Presented here is a scheme for recovery upon failure of a mobile host.
This report portrays the limitations of the mobile wireless environment and
improvement over existing recovery scheme is presented which suit the mobile
environment.

ISE Page 2 DSCE


INTRODUCTION

The portable computers are fast catching up with the desktop workstations in
terms of computing power, memory, display and disk storage. This gives users
the information accessing capability regardless of the location.
Telecommunication companies are continuously improving the quality,
security, availability, and reliability of cellular services and trying to enhance
its scope by providing new services like data management capabilities, which is
highly desirable. Motivated by such growing demand, an information processing
system is referred as the Mobile Database system (MDS). Mobile Database
System (MDS), in essence is a distributed client/server database system where
the entire processing environment is mobile. The actual database may be static
and stored at multiple sites but the data processing nodes, such as laptop,
PDA, cell phones, etc., may be mobile and can access desired data to process
transactions from anywhere and at any time. In this report the main aim i s to
provide a scheme which will help a Mobile Unit (MU) recover efficiently,
taking reduced amount of time to recover from its last secure state. In the
process a strategy for efficiently managing a unified log for a MU is provided.

ISE Page 3 DSCE


PREVIOUS WORK

 LAZY SCHEMES:- Lazy and Pessimistic schemes (asynchronous
schemes) are reported in [8]. In a lazy scheme, logs are stored in the
base station (BS) and, if the MU moves to a new BS, a pointer to the old
BS is stored in the new BS. Pointers can be used during failure to
recover the log distributed over several BS. This scheme has the
advantage that it incurs relatively less network overhead during handoff
as no log information needs to be transferred. Unfortunately, this
scheme has a large recovery time.
 PESSIMISTIC SCHEME:- In the pessimistic scheme, the entire log
and checkpoint record, if any, are transferred at each handoff. Hence,
the recovery is fast but each handoff requires large volume of data
transfer.
 SCHEMES BASED ON THE MU’S MOVEMENT:- In these
schemes, the list of BSs where the log is distributed is transferred during
a handoff. This schemes fail when a MU recovers in a BS different than
the one in which it crashed.
 Distance based:- In the distance-based scheme, log
unification is done when the distance covered by MU
increases above a predefined value.

 Frequency based :- In the frequency-based scheme, log
unification is performed when the number of handoffs seen
by the MU increases above a predefined value.

ISE Page 4 DSCE


SYSTEM MODEL

Mobile networks have two components, a fixed backbone and a wireless
network. There are two distinct sets of entities, namely, mobile hosts and fixed
hosts. A host that can move while retaining its network connection is a mobile
host. The static network comprises of the fixed hosts and the communication
links between them. Some of the fixed hosts, called base stations (BS) are
augmented with a wireless interface, and, they provide a gateway for
communication between the wireless network and the static network.

Due to the limited range of the wireless transceivers, a MU can communicate
with a BS only within a limited geographical region around it. This region is
referred to as a Cell. The geographical area covered by a cell is a function of
the medium used for wireless communication. Due to mobility, the MU may
cross the boundary between two cells while being active. Thus, the task of
forwarding data between the static network and the mobile host must be
transferred to the new cell's BS. This process is known as Handoff.

ISE Page 5 DSCE


PROBLEM SPECIFICATION
MDS recovery process is significantly more complex than conventional
systems mainly because of the following unique aspects of Mobile Environment

 LOCATION IS NOT FIXED:- As the user moves from one place to
another, the location of the MU in the network changes.
 DISCONNECTION:- A MU can get disconnected. While disconnected,
the mobile host is not able to send or receive any message. Protocols
that need to exchange messages will not work correctly in this situation.
During disconnection, the checkpoint protocol should provide a local
recovery mechanism that allows the MU to recover from its own failures.
 BATTERIES HAVE LIMITED POWER:- The MU is often powered by
batteries. Network transmissions and disk accesses are the two major
power consuming operations. To minimize power consumption, the
checkpoint protocol should reduce the amount of information that it
adds to the application messages, and it should avoid sending extra
messages.
 DIFFERENT TYPES OF FAILURES:- MU failure can be separated
into two different categories.
 Hard Failure:- It includes all failures that can not be
repaired; for example, the MU falls and breaks, or is lost
or stolen.
 Soft Failure:- It does not permanently damage the MU; for
example, the battery is discharged and the memory
contents are lost, or the operating system crashes.

ISE Page 6 DSCE


 RANDOM HANDOFF :- MU’s may be subjected to handoff randomly.
A handoff may affect recovery mainly because the location of the desired
MU may not be immediately available for communication.
 WEAK WIRELESS LINK :- The BS and the MU are connected with an
unreliable wireless link which is usually slow.

ISE Page 7 DSCE


METHODOLOGY

Recovery is a mechanism to restore an intermediate consistent state of the
system which may be useful to avoid loosing all the work performed. Recovery
is made possible by maintaining log. Thus, an efficient recovery scheme
requires that the log management should be very efficient and must consume
minimum system resources and recreate the execution environment as soon
as possible after MU reboots.

The entire log must be carefully managed to help MU recover from every
possible failure in minimum time. Now Depending on the system specifications
and requirements, the appropriate recovery and handoff strategy needs to be
chosen. There are several factors which affect the recovery:-

 FAILURE RATE OF THE HOST:- System failures are caused by
defects introduced in manufacturing or by transient or permanent faults
occurring during operation.

 COMMUNICATION/ MOBILITY RATIO:- Communication refers to
the number of messages sent/received by the mobile host and mobility
refers to the number of moves the MU makes in a given period of time.
 MESSAGE SIZE:- Transmitting and receiving data consumes
additional power. In general, transmitting a given amount of data
consumes twice as much power as receiving the same amount of data.
 MEMORY CONSTRAINTS:- A BS generally has many MUs in its cell. It
is necessary to evaluate average memory requirements based on statistical
data and the recovery schemes used.

ISE Page 8 DSCE


 RECOVERY TIME: This is essentially the time required to recover a
process upon failure. If the process has hard real-time deadlines, or
requires high availability, the recovery upon a failure should be quick.

ISE Page 9 DSCE


PROPOSED RECOVERY SCHEME

In our approach we concentrate on recovering the MU state on its failure at the
time of handoff. A region in our approach is a set of all BS controlled by a
Mobile Switching Center. Designated Base Station(DBS) is the station where a
MU registers itself for first time within a Region. It has following
characteristics:-

1. All log activity performed by a MU will be ultimately collected at this DBS

2. In every Region MU will have a different DBS.

Concept of Designated Base Station

INTER-REGION STRATEGY:- Amongst the region the approach works in
a way so as to avoid extra communication steps when MU moves from one
region to a new region. In this case MU has some log maintained at DBS in
Region1, now when the MU moves to Region 2, the first BS with which MH
registers becomes its DBS for new Region say Region 2. So Log will be
transferred from Region1 to Region2.

ISE Page DSCE
10


Region (R)1 Region (R)2

Inter Region Strategy

INTRA-REGION STRATEGY

 LOG TRANSFER FROM MOBILE CACHE TO BS:- When should
MU transfer the entire gathered log to the BS. There are
two cases, first of all if suppose MU’s cache is exhausted,
immediately entire Log will be copied to current BS. In
other case whenever MU moves away from the current BS
and system detects Handoff, the MU will copy the entire log
to BS and will move to new cell and the log would be
appended to the previous Log file.

Log appended

Log Transfer from Mobile Cache to Base Station

ISE Page DSCE
11


 LOG MANAGEMENT:- MU carries following information for the
purpose of registration-

1. Previous BS identity.
2. DBS identity.
3. Own identity in the form of Connect (MHid, PBSid, DBSid).
When any MU registers with a new BS which is not the DBS then the
new BS say BS1 informs DBS about its reachability, saying ‘MU is with
me’. Now since this message is received by DBS so it won’t do anything,
just discard that message. The sequence of messages is shown in the
figure below

Case 1 of Log Management
Now when MU moves from BS1 to BS2, then BS2 informs BS1 about
its reachability. After receiving this info BS1 won’t pass delete
message to DBS as BS1 knows that previous BS for the MU is DBS
where all Log is collected. Now when MU moves further to another BS
after handoff, say BS3 then again the same procedure is repeated. BS3
will inform BS2 about MU’s reachability. Then BS2 will send delete
message to BS1.After receiving the Delete message BS1 copies the
entire log for MU which exists locally over to DBS and after it gets
acknowledgement for data transfer it deletes that log entry from local
space.
ISE Page DSCE
12


Now if MU moves back to BS2 from BS3 then:-
1. BS2 informs BS3 about MU’s reachability.
2. BS3 checks if MU’s previous BS i.e. BS2 and its current BS are same.
If both the BS are same then:
- BS3 won’t send delete message to BS2.
- BS3 will append Log to BS2 log.
- BS3 will delete its copy of log.
- BS2 clears its previous BS identity.
- From BS2 MU will continue with original chain of activities.


ISE Page DSCE
13


Depending on where the MU recovers, there exist two cases for recovery:
Case 1: If MU recovers in same BS where it crashed, then, that MU can
use Log stored at that BS to rollback to previous state. If that log is not
sufficient, then MU will need Log from its Previous BS. If even that is
also not enough, then BS will contact MU’s DBS to restore MU with all
its previous state.
Case 2: If MU fails in one BS, goes to another BS, and recovers there,
then Handoff does not happen and also the new BS does not know
about Previous BS identity. In this case the BS in which MU fails
notifies DBS about the same and whenever MU tries to recover, the new
BS fires query to DBS asking about the log for MU.


ISE Page DSCE
14



ISE Page DSCE
15


MERITS AND OF PROPOSED SCHEME

 Preceding BS log maintenance gives latest recovery option.

 Unification at DBS avoids redundancy of logs.

 Loop is tackled.

 Auto unification of logs at one site helps:

 Quick recovery.

 Achieve real time system deadlines.

 Didn’t require to traverse a link list

DEMERITS OF PROPOSED SCHEME

 DBS with many MUs gets overburdened.
 Single point of failure.

ISE Page DSCE
16


FUTURE ASPECTS

For more efficient management of recovery information and faster recovery,
some other schemes can be employed which uses checkpoints and message
logs together in order to guarantee recovery and overcome the cost of storing
entire log at one place. The movement- based scheme can also be used which
considers Mobile Switching Center as a stable storage which considers both
of the failure-free operation cost and the recovery cost. Maintaining entire log
at one place in long run may make BS to run out of space so strategies should
be developed for Garbage Collection to flush the log at regular intervals.

ISE Page DSCE
17


CONCLUSION

In this paper, an efficient recovery scheme based on message logging for mobile
computing systems is presented. The mobile host carrying its recovery
information to its current BS can recover instantly in case of a failure. In the
proposed approach, the recovery information is not dispersed over a wide range
of cells so the recovery cost is not too high. As described in the scheme, while
the mobile host moves within a region, recovery information of the mobile host
is transferred among the BS of that region. However, if the mobile host moves
out of the region, then DBS transfers the recovery information to another
region’s DBS. The scheme controls the transfer cost as well as the recovery
cost.

ISE Page DSCE
18


REFERENCES

[1] Nuno Neves, W. Kent Fuchs, “Adaptive Recovery for Mobile Environments”,
ACM Press, Vol.40, Issue 1, pp. 68 – 74, 1997.

[2] Bin Yao, W. Kent Fuchs, Kuo-Feng Ssu. "Message Logging in Mobile
Computing", IEEE Twenty-Ninth Annual International Symposium on Fault-
Tolerant Computing, 1999.

[3] Cris Pedregal Martin, Krithi Ramamritham, “Recovery Guarantees in
Mobile Systems”, Proc. of the 1st ACM international workshop on Data
Engineering for Wireless and Mobile Access, pp. 22-28, 1999.

[4] Cris Pedregal-Martin, Krithi Ramamritham, “Support for Recovery in
Mobile Systems”, IEEE Transactions on Computers, v.51 no.10, pp. 1219-
1224, October 2002.
[5] Dhiraj K, Pradhan, P. Krishna, Nitin H. Vaidya, “Recoverable Mobile
Environments: Design and Trade-off Analysis”, ACM Press, 2001.
[6] Taesoon Park, Namyoon Woo, Heon Y. Yeom, “An Efficient Recovery
Scheme for Mobile Computing Environment”, ICPADS.
[7] Taesoon Park, Namyoon Woo, Heon Y. Yeom,“An Efficient Optimistic
Message Logging Scheme for the Recoverable Mobile Computing Systems”,
ISAS-SCI (1), 2001.
[8] P. Krishna, N. H. Vaidya and D. K. Pradhan, "Recovery in Distributed Mobile
Environments", IEEE Workshop on Advances in Parallel and Distributed
Systems, 1993.
[9] Sashidhar Gadiraju, Vijay Kumar, “Recovery in the Mobile Wireless
Environment Using Mobile Agents”, IEEE Transaction on Mobile Computing,
Vol. 3, No. 2, pp. 180-191.
[10] R. Koo, and S. Toueg,“Checkpointing and Rollback-Recovery for
Distributed Systems”,IEEE Trans. Software Eng.,Vol.13,No.1, pp. 23-31, 1987.

ISE Page DSCE
19


[11] L. Alvisi, and K. Marzullo, “Message Logging: Pessimistic, Optimistic,
Causal, and Optimal”, IEEE Trans. Software Eng., Vol. 24, No. 2, pp. 149-159,
1998.
[12] H. Higaki, M. Takizawa, “Checkpoint Recovery Protocol for Reliable Mobile
Systems”, the 17th IEEE Symposium on Reliable Distributed Systems.
[13]CHENG-MIN LIN AND CHYI-REN DOW, “Efficient Checkpoint-based Failure
Recovery Techniques in Mobile Computing Systems”, Journal of information
science and engineering 17, pp. 549-573, 2001.
[14] T. Park, N. Woo, H. Yeom, “Efficient Recovery Information Management
Schemes for the Fault Tolerant Mobile Computing Systems”, 20th IEEE
Symposium on Reliable Distributed Systems (SRDS'01).

ISE Page DSCE
20


not even use the network's internet connection for Web surfing or other online
activities, this is referred to as piggybacking. [2] The idea behind this is to just
find the network and see that it exist, but don't connect to it in any way. This
will prevent you from breaking the law, because connecting to it is stealing and
is illegal by today's laws. Just because as a wardriving you don't mean to
connect doesn't mean that it is illegal. This can be a problem with some client
adapters that auto connect to the a network whenever it comes into range. This
is still illegal to connect to any of there non-public access points without
permission. The keys on this part is a wardriver is responsible for the hardware
that they choose, so make sure that it is running as expected and don't get
caught in a situation where something illegal could be taking place. The main
thing with the legal issues here is that wardriving is legal and can be beneficial,
but this does not mean that everything a wardriver does can be ethical.
Wardrivers should take the knowledge they gain from doing such task and help
notify people of their vulnerabilities. Having all this information on networks in
a given area brings into some ethical issues that a wardriver must make sure
they follow, because a wardriver has a database of all these vulnerable
networks, it makes it easy for them to get caught into ethical dilemmas. This
information can be both used for good and for bad. The good notifying people of
their vulnerabilities and the bad taking advantage of these networks or selling
information about them to others who might take advantage of this
information. This is each users responsibility to ask themselves, is what I'm
doing ethical. If the answer is no or even shady it is time to stop doing it and

ISE Page DSCE
21


think about the purpose of wardriving and the effects each wardriver has on
the technology community.

WARDRIVING METHODOLOGY

THE RIG

A “rig” refers to all of the hardware and software components used while
wardriving. The forum is a valuable resource, especially to new wardrivers,
because they may learn what components are effective and not effective for
wardriving. This forum also allows security professionals to observe how the
activity i s progressing over time and may even assist in recognizing a
wardriver on the road.

COMPUTER SYSTEM

The foundation of every wardriving rig is a computer system. Most wardrivers
choose to use either a laptop or PDA since these devices are relatively light and
portable and they may function on batteries if necessary. A simple power
inverter may be used with the cigarette lighter adaptor found in most vehicles
to provide power for the entire rig.

The selection of a laptop or PDA will affect your available scanning software
options. An x86-based laptop may utilize software on several Linux and Wi
ndows operating systems. A PowerPC-based laptop will be able to uti lize
software on Mac OS X or Yellow Dog Linux. W ith a Pocket PC PDA, one would
be restricted to software that runs on Familiar Linux or Microsoft Pocket PC,
while Palm OS devices do not have any relevant software availabl e for
wardriving at the moment.

ISE Page DSCE
22


ISE Page DSCE
23

report on recovery in mobile database system

Recommended

Recommended

More Related Content

What's hot

What's hot (18)

Similar to report on recovery in mobile database system

Similar to report on recovery in mobile database system (20)

Recently uploaded

Recently uploaded (20)

report on recovery in mobile database system