new idea

OVERVIEW OF THE PROJECT
A mailing service system will be developed by us which will be benefcial for
the visually impaired. Here, the user will have to use certain keywords which
will perform certain actions.
The keywords are:
List
ead
!top
"ompose
CONTINUED
A recorded voice will guide the user throughout the process.
The mail composed by the user via voice will be converted to te#t before
sending it to a normal individual.
EXPECTED OUTCOME
!peech to te#t conversion.
Te#t to speech conversion.
Allowing auto correction of mail by reading it aloud before sending.
A handy device to carry out these operations.
THA$% &'(
)'*A+$, HA)-A.
'/.AT+$0 !&!T.*
'-..$A AL1A 2 34
5'$+TA 6.$A$).! 2 78
A$$ L+!A 5'!./H , 9:
)esigning a 1oice *ail system architecture that can be used by a blind person
to access mails easily and e;ciently.
Thus reliance of visually impaired on other people for their activities related to
mail can be reduced.
SCOPE OF THE PROJECT
PRE-REQUISITES
TECHNOLOGIES
eal time composition of te#tual mail through audio input.
(se of keywords to trigger particular mail related operations.
The user can scan the received mails and also save a softcopy of the mail.
Audio output of recieved mail.
A handy device with minimal user interface which reduces the reliance on
sighted people.
"oncept of te#t to speech conversion.
!*T/<!imple mail transmission protocol for mail servicing=.
%nowledge of hardware re>uirements for te#t to speech.
"*(!phin#: open source software for speech to te#t.
e!peak : open source software for te#t to speech.
REFERNCES
? 1oice @ased !ystem in )esktop and *obile )evices for @lind /eople ?,
+!!$ 9943,9:48, +!' "ertifed 5ournal, 1olume :, +ssue 9,
6ebruary 937:.
? Te#t to !peech "onversion with /honematic "oncatenation?, +nternational
5ournal of .lectronics "ommunication and "omputer Technology <+5.""T=
1olume 9 +ssue 4 <!eptember 9379=.
Cloud Computing concerns large-scale interconnected systems and it has the main purpose to aggregate and to efcient
exploit the power of widely distributed resources. Resource Management and Task Scheduling play an essential role, in
cases where one is concerned with optimized use of these resources. The ubiquitous networks are highly dynamic
distributed systems so the changes in overlay are frequent. On the other hand, the Cloud systems are highly dynamic in
its structure because the user requests must be respected as an agreement rule. When ubiquitous networks become
clients for Cloud systems new algorithm for events and tasks scheduling and new methods for resource management
should be designed in order to increase the performance of such systems. The adaptive methods used in context are
oriented on: self-stabilizing, self-organizing and autonomic systems; dynamic, adaptive and machine learning based
distributed algorithms; fault tolerance, reliability, availability of distributed systems.
A major goal of this special issue is to explore new directions and approaches for reasoning about resource management
in Cloud Computing based on adaptive methods, and to encourage the submission of ongoing work with already
important theoretical and practical results, as well as position papers and case studies of existing verifcation projects to
highlight the art in this domain. We expect that this special issue will draw signifcant number of submissions, readership,
and consequently numerous citations of the published contributions.
TM Forum defines SLAs as expectations among two or more parties regarding service quality, priorities, and
responsibilities. Te !loud Standards !ustomer !ouncil views cloud SLAs as written expectations for service
between cloud consumers and providers. "t provides guidance to decision ma#ers on wat to expect and wat
to be aware of as tey evaluate and compare end user SLAs from cloud computing providers. Te decision
ma#ers sould also evaluate te SLAs tat a cloud computing provider as wit vendors, enterprise data
centers, networ# providers, and content providers.
An SLA is not a mandate wen it$s driven by a ma%or reorgani&ation, downsi&ing, service consolidation, or
transition to a cloud services environment. "t doesn$t ave te inputs from all pertinent parties tat must be
involved.
An SLA is not a one'way solution. (ne party ) te cloud service provider, for example ) sould not impose
decisions about ow tings sould be done, particularly wen te oter party ) te cloud service customer, for
example ) as different expectations about ow te SLA sould be formulated.
An SLA isn$t a quic# solution. *using te process of negotiating te terms and conditions in te SLA doesn$t
give enoug time for te parties to understand eac oter$s expectations particularly wen eac party as a
different perception of wat a certain terminology stands for.
AuthorAs real,life story on !LA
A couple of years ago, developer+or#s publised a seven'part article series " wrote about SLAs
,see *esources-. Afterwards, " was contacted by te cief of web services and services oriented arcitecture for
a ma%or .nited States federal agency. /e particularly li#ed my article, .se SLAs in a +eb services context, 0art
12 Mitigate ris# for vulnerability wit a SLA guarantee. Te article tal#s about interruption tresolds tat can
spell te difference between saving or not saving an enterprise from a complete 3enial of Service ,3oS- and
destruction of critical system assets, and ow large interruption tresolds can impact te SLA guarantees.
+ile corresponding wit te agency cief, " mentioned anoter article in tat series, .se SLAs in a +eb
services context, 0art 42 5uarantee your +eb service wit a SLA. "n tat article, " tal# about six testing types
you sould use to test web services features before a service is made public2
• Statefulness2 3oes te server respond in te correct steps6
• Access2 !an an unautori&ed user successfully access a control tat only administrators are autori&ed to use6
• *esponse time2 "s te application ta#ing too long to respond6
• Time'out2 +at appens wen te application times out6
• 7ersioning2 !an a new build brea# an existing application$s functions6
" also mention te exceptions tat potentially can be included in te SLA2
• Failure2 /ardware, telecommunications, software, performance monitor.
• 8etwor# issues not witin direct control of service provider.
• 3enial of service2 !lient negligence9willful conduct: Acts of 5od: war stri#es, unavailability of
telecommunications: inability to get supplies or equipment needed for te provisions of te SLA.
• Sceduled maintenance2 /ardware and software upgrades and bac#ups.
Te agency cief as#ed me to do a presentation and old a systems engineering discussion wit is team. "n
my presentation, " included all te points above, plus " added te importance of an exit clause in te SLA. Tis
clause allows te customers to exit from te SLA wen te guarantees can$t be met at numerous times.
After my presentation was reviewed, te cief sent me a copy of is agency$s SLA template for review and
as#ed me to contribute to te first and later versions of te template. ,0rivacy reasons prevent me from
discussing details ere.- "t was eventually made part of te federal system.
Tis SLA template can be extended to cloud'based services. All SLAs are about service guarantees regardless
of te type of service provided2 +eb services, cloud services, and networ# access services. 0enalties are
imposed if service guarantees are not met. Te level of guaranteed services differs from one partner to anoter.
8ot all web services ,suc as S(A cloud'based applications- are cloud'based. S(A is a design pattern
composed of loosely coupled, discoverable, reusable, inter'operable, platform'agnostic services via web
services standards. SaaS is not S(A. SaaS is a consumption model: it consumes resources osted by a cloud
service provider. S(A is a design model in wic tere is no restriction on wo te consumer is.
Back to top
-hy itAs important to standardiBe !LAs in multi,cloud environment
+ile SLAs ave traditionally been a contract between a service provider and a cloud service customer ,an
enterprise, business, or government agency-, te expanding value cain for oter services as made SLAs
important for a myriad of often complex relationsips between partnersips.
For example, te same service provider can provide services to2
• !loud service customers ,SaaS end users, 0aaS developers and "aaS infrastructure specialists-
• 7endors
• Large enterprises
• ;usinesses
• 5overnment agencies
Te same vendor provides services to2
• 8etwor# providers
• !loud service providers
• +eb service providers
• <nterprises
• ;usinesses
• 5overnment agencies
Te same networ# provider provides networ# access services to cloud service providers.
Te same enterprise provides private cloud services to SaaS end users, 0aaS developers, and "aaS
infrastructure specialists.
As part of pre'disaster planning, a requesting provider needs to discover if reserved resources would be
available from a substitute provider. "t can be a many'to'many relationsip between requesting and substitute
providers. A requesting provider can ave more tan one substitute provider. A substitute provider can
associate wit one or more requesting providers.
Te cloud service provider is te substitute cloud service provider wo accepts te request for data transfer
services from a requesting cloud service provider so tat te requesting provider can fix te problem wit
outage or load balancing. Te substitute provider can limit te number of data transfer requests from more tan
one requesting service providers.
To compete successfully, companies must proactively manage te quality of teir services. Since provisioning
of tose services is dependent on multiple partners, management of SLAs become critical for success. More
important is te standardi&ation of terminologies in te SLAs among te partners.
+en terminologies are not standardi&ed, te definition of a terminology tat a partner ,a requesting service
provider- uses may differ from te definition tat anoter partner ,a substitute service provider or vendor- uses.
0artners, for instance, may define SLA parameters ,suc as failure to meet performance- in different ways.
+en te differences aren$t ironed out during a negotiation, tey migt impact te SLA process ,for example,
wat is a penalty and wen to impose te penalty-. (ne partner migt impose te penalty of a 4= percent
rebate of service fees for cloud service downtime exceeding one our wile anoter partner migt impose te
penalty of 4> percent for downtime exceeding alf an our.
Back to top
!LA best practices
;est practices are used to describe te process of developing and following a standard way of doing tings tat
multiple partners can use. +en te partners agree on te standard way of using terminologies wen
negotiating SLAs, tey contribute to te process of SLA standardi&ation.
Te !loud Standards !ustom !ouncil provides cloud consumers wit te seven steps tey sould ta#e wen
evaluating cloud SLAs to elp tem #now wat to expect wen comparing cloud service providers or
negotiating terms wit a provider, as detailed in 0ractical 5uide to !loud Service Level Agreements, publised
by te !loud Standards !ustomer !ouncil in April >=4> ,see *esources.
!tep 7: +dentify the cloud actors
Te first step is to identify te cloud actors. Te 8ational "nstitute of Standards and Tecnology ,8"ST-
*eference Arcitecture ,see *esources- identifies five unique cloud actors2
4. !loud consumer
>. !loud provider
?. !loud carrier
@. !loud bro#er
A. !loud auditor
<ac actor as a unique role and responsibilities. (nly te first tree ave relationsips wit one anoter
regarding te terms and conditions in te SLAs. (ne drawbac# to note is tat te 8"ST *eference Arcitecture
doesn$t identify oter actors suc as cloud vendor and web services providers. Anoter drawbac# is tat it
doesn$t say if tese actors ave many'to'many relationsips.
!tep 9: .valuate business level policies
Te policies expressed in te SLA sould be evaluated against te business strategy and policies. Te data
policies tat consumers need to consider for inclusion in te cloud SLA wen reviewing a cloud SLA are data
preservation, redundancy, location, sei&ure and privacy.
Te business level policies tat sould be considered for inclusion in te SLA include guarantees, list of
services not covered, excess usage, payment and penalty metods, subcontracted services, licensed software
and industry specific standards.
!tep C: (nderstand !aa!, /aa! and +aa!
Te tird step is to understand wat SaaS, 0aaS, and "aaS are about and wic type of cloud it is running on
,private public or ybrid-. Terms and conditions in te SLA depend on te complexity of control variables tat
te provider gives to te consumers.
Te SLA for te SaaS is not as complex as te SLA for te 0aaS. Te only control te SaaS consumer ,end
users- as is to access te SaaS application, wile te 0aaS consumers ,developers- ave controls over te
application development life cycle but not te virtual macines. Te SLA for te "aaS is te most complex as te
"aaS consumers ,infrastructure specialists- ave control over te virtual macines but not pysical
infrastructure.
!tep :: *etrics
Te fourt step is to identify wat metrics sould be used to acieve performance ob%ectives. Some examples
of availability and response time metrics are2
• Metric name in SLA
• !onstraints
• Metod and frequency of collection
!tep 4: !ecurity
!onsider #ey security requirements for cloud SLAs, including2
• Asset sensitivity
• Legal9regulatory requirements
• !loud providers$ security capabilities
<ac country as a different set of privacy regulations tan anoter country. For tis reason, te consumers
sould #now in wic country te data would be stored in te cloud. (ne country may proibit certain privacy
data from outside te country, wile anoter country may allow external privacy data.
!tep D: +dentify service management re>uirements
"dentify service management requirements. Tey include wat sould be monitored and reported ,for example,
load performance, application performance-, and wat sould be metered. Tey also include ow rapid
provisioning sould be ,speed, testing, demand flexibility- and ow resource cange sould be managed.
!tep E: /repare for and manage service failure
Te final step is to prepare for and manage service failure, determining wat remedies sould be provided ,for
example, service credits- and wat are te liability limitations.
Ten you need to understand ow te disaster recovery plan will wor# wen needed. Te plan sould define
wat service outage is, ow unexpected incidents will be andled, and wat actions to ta#e wen service
disruption is prolonged.
An exit clause sould be part of every cloud SLA in case eiter te consumer or provider wants to terminate te
relationsip.
Back to top
"loud metrics
"n te S0<! (pen Systems 5roup, te !loud !omputing +or#ing 5roup wrote a B*eport on !loud !omputing
to te (S5 Steering !ommitteeB ,see *esources.- Te report provides a list of cloud metrics used to measure
ow well te tests are performing.
Te list includes2
• <lasticity ,provisioning internal, agility, scaling up and down-
• Trougput
• *esponse time
<lasticity metrics include provisioning internal, agility, and scaling up and down. 0rovisioning internal measures
te time needed to bring up or drop a resource. 0rovisioning internal measures te time to bring up2
• A new instance ,resource- on te "aaS
• A new instance of application server on te 0aaS
• A new application instance online to meet increasing demand
Agility measures ow well te wor#load can be scaled and ow well te system is provisioned to be as close to
te needs of te wor#load as possible.
Trougput is te amount of wor# te cloud can do per unit time. *esponse time is te time between wen a
request is made by a user and wen te response is received by te user. (r more specifically ow long does it
ta#e for te application to respond to te user$s request.
Back to top
-hatAs missing from the cloud metrics
Missing from te cloud metrics list are user, data request, and resource tresold levels tat could be used in
determining guaranteed service levels in te SLA.
User threshold level
Sets te maximum number of users concurrently accessing te application tat consists of accepted or
restructured service components decomposed from a legacy system.
Data requests threshold level
Sets te maximum number of data requests tat users can concurrently sent to te SaaS application.
Resources threshold level
Sets te maximum amount of resources ,e.g., !0., storage devices, dis# space- tat can be allocated to eac
SaaS user, 0aaS developer and "aaS networ# specialist.
Cnowing wo sets te tresold levels is important wen te consumers evaluate te SLAs. Setting te
tresold levels depend on te control variables assigned to te SaaS, 0aaS, and "aaS users and providers2
• SaaS user. Te end user doesn$t set any tresold levels. Te only control te user as is to access te
application from eiter te provider or te 0aaS developer from a des#top, laptop, or mobile.
• SaaS provider. Te provider sets te user tresold level. At a minimum, te provider controls operating
systems, ardware, networ# infrastructure, SaaS application upgrades, and patces.
• 0aaS application developer. Te developer can set te data request tresold at te SaaS user level for SaaS
users wo are co'residents on te 0aaS. Te developer controls te development of all SaaS applications of
loosely coupled service components wit accepted dependencies in a full business life cycle, and runs tem on
te 0aaS.
• 0aaS provider. Te provider sets te data request and resource tresolds at te developer level. At a
minimum, te provider controls operating systems, ardware, networ# infrastructure, and resource
management.
• "aaS infrastructure and networ# specialist. "f te specialist osts te 0aaS, e may set te user and data
requests for te 0aaS developers. Te specialist controls te operating systems, networ# equipment, and
deployed applications at te virtual macine level.
• "aaS provider. Te provider sets user, data requests, and resource tresold levels. At a minimum, te provider
controls te infrastructure of traditional computing resources in te cloud environment.
Back to top
+n conclusion
"n planning for SLA standardi&ation, consider best practices for developing and following a standard way of
doing tings tat multiple partners can use. 8ew terminologies suc as user, data request, and resource
tresolds sould be considered for use in determining guaranteed service levels. +en te partners agree on
te standard way of using terminologies wen negotiating SLAs, tey contribute to te process of SLA
standardi&ation. +e need to build a team of developers, managers, business analysts, system engineers and
ma#e it easier for tem to teir %ob of standardi&ing te terms and conditions in te SLA for eac deployment
type2 SaaS, 0aaS and "aaS.
Resources
Learn
• 7isit TM Forum, a global, non'profit industry association focused on enabling service
provider agility and innovation.
• Te !loud Standards !ustomer !ouncil is an end user advocacy group dedicated to
te accelerating cloud$s successful adoption, and drilling down into te standards,
security and interoperability issues surrounding te transition to te cloud.
• See te articles te autor refers to about SLAs for web services2 .se SLAs in a +eb
services context, 0art 42 5uarantee your +eb service wit a SLA, and .se SLAs in a
+eb services context, 0art 12 Mitigate ris# for vulnerability wit a SLA guarantee.
• Te 0ractical 5uide for Service Level Agreements, publised by te !loud Standards
!ustom !ouncil ,!S!!-, igligts te critical elements of a service level agreement
for cloud computing and provides guidance on wat to expect and wat to be aware of
wen negotiating an SLA.
• Te 8ational "nstitute of Standards and Tecnology ,8"ST- *eference
Arcitecture describes an overall framewor# for use in cloud computing for te ..S.
Federal 5overnment.
• *ead more about cloud metrics in te *eport on !loud !omputing to te (S5 Steering
!ommittee, written by te !loud !omputing +or#ing in te Spec (pen Systems 5roup.
• Learn more about cloud computing tecnologies at cloud on developer+or#s.
• Follow developer+or#s on Twitter.
• +atc developer+or#s demos ranging from product installation and setup demos for
beginners, to advanced functionality for experienced developers.