disaster recovery
This term is related to Database backup and recovery so it basically used when, any disaster occur then how the recovery is done and what techniques are used.
Content
SYNOPSIS ON DISASTER RECOVERY AND CLOUD SERVICES
Submitted To Lovely Professional University, Phagwara (For the partial fulfillment of the degree of MC !""" S#M$
SUPERVISED BY SUBMITTED BY:
M%& %'("T '(%" ( ssistant Le-turer$
.isaster %e-overy and Cloud Servi-es
)ame * P + ) S"),( %oll )o& * %./#01 21
1
%eg& )o& * //32004/
.ate of Submission* 35 '-t 32/4
ACKNOWLEDGEMENT
+ith profound regard " gratefully a-6nowledge our respe-ted -ourse Le-turer, MR. ROHIT OHRI, .epartment of C , Fa-ulty of -omputer organi7ation and ar-hite-ture, for his generous help and day today suggestion during the preparation of the report& " li6e to give than6s espe-ially to our friends 8 many individuals, for their enthusiasti- en-ouragements and helps during the preparation of this report and for their assistan-e in typing and proofreading this manus-ript&
.isaster %e-overy and Cloud Servi-es
2
Table of Content Contents Page No.
/& -6nowledgement9999999999999999993
3& "nde:99999999999999999999999&4 4& bstra-t99999999999999999999990
0& "ntrodu-tion99999999999999999999&0!; ;& Free 'f Cast Five Tools For .isaster %e-overy99999&&&<!5 <& Servi-es nd Features #ssential For .isaster %e-overy99&&5!= 1& .atabases999999999999999999999= 5& The Cloud "s "deal For .isaster %e-overy&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&=!/2 =& .isaster %e-overy "n The Cloud 'ptions99999999&/2!// /2& >enefits of -loud -omputing9999999999999&// //& Con-lusion99999999999999999999&/3 /3& %eferen-es99999999999999999999&&/3
.isaster %e-overy and Cloud Servi-es
3
ABSTRACT:+ith the growing value of data as a strategi- -orporate asset, today?s "T organi7ations fa-e the -hallenge of implementing reliable ba-6up and re-overy solutions in the most effi-ient, -ost!effe-tive manner& To meet this -hallenge, they need to -arefully define their business re@uirements and re-overy obAe-tives before de-iding on the right ba-6up and re-overy te-hnologies to deploy& "n the event of a disaster, we -an @ui-6ly laun-h resour-es in ma7on +eb Servi-es ( +S$ to ensure business -ontinuity& The paper highlights relevant +S features and servi-es that we -an leverage for our .% pro-esses and shows e:ample s-enarios about how to re-over from disaster& "t further provides re-ommendations about how we -an improve our .% plan and leverage the full potential of +S for our .isaster %e-overy pro-esses&
INTRODUCTION:DISASTER RECOVERY .isaster re-overy, 'perational ba-6up and ar-hival is used in ba-6up purpose& .isaster re-overy is a pro-ess, poli-es and pro-edure that are used to re-over the lost data& Many business and government servi-es utili7e .isaster %e-overy (.%$ systems to minimi7e the downtime raised by -atastrophisystem failures& Currently .isaster %e-overy system use periodi- tape for ba-6up& .isaster re-overy is -lassified into two -ategories, one is natural li6e storms, earth@ua6es et-& and other is made by man li6e ha7ardous materials spills, infrastru-ture failure et-& +hen any disaster re-overy is re@uired, the disaster re-overy planner first report to their organi7ation business -ontinuity plans whi-h should indi-ate the matri-es of %P' (re-overy point obAe-tive$ and %T' (re-overy time obAe-tive$&
.isaster %e-overy and Cloud Servi-es
4
THE NEED FOR BETTER DATA PROTECTION s today?s businesses in-reasingly re-ogni7e -orporate data as a strategiasset that must be prote-ted against a wide range of ris6s and threats, data prote-tion has be-ome a high!priority obAe-tive for "T organi7ations& "n turn, these organi7ations now fa-e the -hallenges of identifying, deploying, and effi-iently managing their data ba-6up and re-overy infrastru-tures& RECOVERY POINT OBJECTIVE AND RECOVERY TIME OBJECTIVE RPO* ! This des-ribes the a--eptable amount of data loss measured in time& For e:ample, if the %P' was / hour, after the system was re-overed, it would -ontain all data up to a point in time that is prior to //*22 M be-ause the disaster o--urred at noon& RTO: - This is the duration of time and the servi-e level to whi-h a business pro-ess must be restored after a disaster (or disruption$ to avoid una--eptable -onse@uen-es asso-iated with a brea6 in business -ontinuity& For e:ample, if a disaster o--urs at /3*22 PM (noon$ and the %T' is 5 hours, the .% pro-ess would ensure re-overy to the a--eptable servi-e level would be possible by 5*22 PM& CLOUD SERVICES >efore des-ribing servi-es of -loud -omputing first we 6now about -loud -omputing& The -loud -omputing is internet based -omputing, where shared all hardware and software resour-es and information to the -omputer (that are -onne-ted to internet$ on demand but paying some -harge& The various servi-es are* B /$ Software as servi-es& 3$ Platform as servi-es& 4$ "nfrastru-ture as servi-es& To minimi7e the impa-t of a disaster on the business, -ompanies invest time and resour-es to plan, prepare, do-ument, train, and update pro-esses to deal with events& The amount of investment for the disaster re-overy planning of a parti-ular system -an vary dramati-ally depending on the -ost of a potential
5
.isaster %e-overy and Cloud Servi-es
outage& This paper des-ribes some typi-al approa-hes ranging from minimal investments to full!s-ale availability and fault toleran-e& Proper preparation for .% is a must, and this paper outlines some of the best pra-ti-es to improve our .% plans and pro-esses& AWS (AMAZON WEB SERVICE) allows us to s-ale up our infrastru-ture on an as!needed basis& +e get a--ess to the same highly s-alable, reliable, se-ure, fast, ine:pensive infrastru-ture that ma7on uses to run its own global networ6 of web sites and only pay for what we use&
FREE OF COST FIVE TOOL FOR DISASTER RECOVERY:The five tools that are for disaster re-overy that are following*! /& Ma-rium %efle-ts& 3& Clone7illa& 4& .rive"mageCML& 0& Dui-6 .isaster %e-overy& ;& System %es-ue C.& 1. MACRIUM REFLECT This is the free version of the Professional produ-t& This version supports CP, Eista, +indows 1, +indows 5 and this is use for home!use des6top& For using this -an easily handle .is6 imaging, s-hedule ba-6up and also supports % ".&
Fig- .1!. !. CLONEZILLA Clone7illa is a free, open sour-e, bare metal ba-6up and re-overy tool& Clone7illa is based on .%>L, Part-lone, and Up-ast& There are two versions of Clone7illa! Live and Clone7illa S# (Server #dition$& The Live version is to be
.isaster %e-overy and Cloud Servi-es
6
used for a single des6top, whereas the Server edition is suitable for massive deployment (up to 02 ma-hine simultaneously$&
Fig- .! ". DRIVEIMA#E$ML .rive"mageCML is similar to Ma-rium %efle-t in that it offers a free version for personal use& This free version allows us to ba-6up, browse, and restores images& +ith the ability to browse images, this means we -an re-over files andFor folder (not for entire image$&
Fig- ."
%.
&UIC' DISASTER RECOVERY
Dui-6 .isaster %e-overy is a tool that -an @ui-6ly re-over a ma-hine when various built!in +indows administrator tools have been disabled (su-h as the %egistry #ditor, Tas6 Manager, et-&$
.isaster %e-overy and Cloud Servi-es
7
Fig- .% (. SYSTEM RESCUE CD System %es-ue C. is a Linu: system res-ue dis6 that allows us to administer and repair a system after a -rash& +e -an manage partitions, re-over data, edit -onfiguration files, and we -an wor6 with both Linu: and +indows systems&
Fig- .(
SERVICES
AND
FEATURES
ESSENTIAL
FOR
DISASTER RECOVERY:-
.isaster %e-overy and Cloud Servi-es
8
"n the preparation phase of .%, it is essential to -onsider the use of servi-es and features that support data migration and durable storage be-ause they enable we to restore ba-6ed up -riti-al data to +S when disaster stri6es& ma7on +eb Servi-es ( +S$ is the first Cloud vendor that 'ra-le has partnered with to enable database ba-6up in the Cloud& RE#IONS ma7on +eb Servi-es are available in multiple %egions, so we -an -hoose the most appropriate lo-ation for our disaster re-overy site& STORA#E ma7on Simple Storage Servi-e ( ma7on S4$ provides a highly durable storage infrastru-ture designed for mission -riti-al and primary data storage& COMPUTE ma7on #lasti- Compute Cloud ( ma7on #C3$ provides resi7able -ompute -apa-ity in the -loud& +ithin minutes, we -an -reate #C3 instan-es, whi-h are virtual ma-hines over whi-h we have -omplete -ontrol& NETWOR'IN# +hen dealing with a disaster, it?s very li6ely that we will have to modify networ6 settings as we are failing over to another site& A)*+,- R,./0 ;4 is a highly available and s-alable .omain )ame System (.)S$ web servi-e& "t is designed to give developers and businesses an e:tremely reliable and -ost!effe-tive way to route end users to "nternet appli-ations& ma7on .ire-t Conne-t ma6es it easy to set up a dedi-ated networ6 -onne-tion from our premise to +S&
DATABASES:For our database needs, -onsider using these +S servi-es* A)*+,- R01*/i,-*1 D*/*2*30 S045i60 (A)*+,- RDS) ma6es it easy to set up, operate, and s-ale a relational database in the -loud& Could use ma7on
.isaster %e-overy and Cloud Servi-es
9
%.S either in the preparation phase for .% (.isaster %e-overy$ to hold our -riti-al data in a running database& A)*+,- Si)710 DB is a highly available, fle:ible, non!relational data store that offloads the wor6 of database administration& "t -an also be used in the preparation and the re-overy phase of .%&
THE CLOUD IS IDEAL FOR DISASTER RECOVERY:Disaster recovery in the cloud is a relatively new concept, and like many technology trends, there's a lot of hype and misinformation out there. Cloud infrastru-ture is gaining in popularity for organi7ations that have very real business needs that rely on information te-hnology ("T$ infrastru-ture but struggle with the -osts both -apital and operational of e:panding their data -enters& >ut before even dis-ussing the -loud, it?s worthwhile to settle on a definition& .% is an ideal use -ase for ta6ing advantage of the -loud& +hile many organi7ations remain -autious about pla-ing produ-tion servi-es in the -loud, they are often more -omfortable testing those waters for .%Gespe-ially sin-e the -loud alters the e-onomi-s of .% so radi-ally& >ut disaster re-overy in the -loud isn?t a perfe-t solution, and its short-omings and -hallenges need to be -learly understood before a firm ventures into it& Se-urity usually tops the list of -on-erns*
I3 8*/* 306.4019 /4*-3:04408 *-8 3/,408 i- /;0 61,.8< H,= *40 .3043 *./;0-/i6*/08< A40 7*33=,483 /;0 ,-19 ,7/i,- ,4 8,03 /;0 61,.8 74,5i804 ,::04 3,)0 /970 ,: /=,-:*6/,4 *./;0-/i6*/i,-< D,03 /;0 61,.8 74,5i804 )00/ 40g.1*/,49 40>.i40)0-/3<
nd be-ause -louds are a--essed via the "nternet, bandwidth re@uirements also need to be -learly understood& There?s a ris6 of only planning for bandwidth
10
.isaster %e-overy and Cloud Servi-es
re@uirements to move data into the -loud without suffi-ient analysis of how to ma6e the data a--essible when a disaster stri6es*
• •
D, =0 ;*50 /;0 2*-8=i8/; *-8 -0/=,4? 6*7*6i/9 /, 408i406/ *11 .3043 /, /;0 61,.8< I: =0 71*- /, 403/,40 :4,) /;0 61,.8 /, ,--740)i303 i-:4*3/4.6/.40@ ;,= 1,-g =i11 /;*/ 403/,40 /*?0<
DISASTER RECOVERY IN THE CLOUD OPTIONS:MANA#ED APPLICATIONS AND MANA#ED DR n in-reasingly popular option is to put both primary produ-tion and disaster re-overy instan-es into the -loud and have both handled by a )*-*g08 3045i60 74,5i804 (MSP$& >y doing this we?re reaping all the benefits of -loud -omputing, from usage!based -ost to eliminating on!premises infrastru-ture& "nstead of doing it our self, we?re deferring .% to the -loud or managed servi-e provider& BAC' UP TO AND RESTORE FROM THE CLOUD ppli-ations and data remain on!premises in this approa-h, with data being ba-6ed up into the -loud and restored onto on!premises hardware when a disaster o--urs& "n other words, the 2*6?.7 i- /;0 61,.8 be-omes a substitute for tape!based off!site ba-6ups& +hen -ontemplating -loud ba-6up and re-overy, it?s -ru-ial to -learly understand both the ba-6up and the more problemati- restore aspe-ts& >a-6ing up into the -loud is relatively straightforward, and ba-6up appli-ation vendors have been e:tending their ba-6up suites with options to dire-tly ba-6 up to 7,7.1*4 61,.8 3045i60 74,5i8043 3.6; *3 ATAT@ A)*+,-@ Mi64,3,:/ C,47,4*/i,-3. The -hallenging aspe-t of using -loud!based ba-6ups for disaster re-overy is the re-overy& +ith bandwidth limited and possibly terabytes of data to be re-overed, getting data restored ba-6 on!premises within defined %T's -an be -hallenging&
.isaster %e-overy and Cloud Servi-es
11
NEW OPTIONS@ OLD FUNDAMENTALS The -loud greatly e:tends disaster re-overy options, yields signifi-ant -ost savings& "t does not, however, -hange the .% fundamentals of having to devise a solid disaster re-overy plan, testing it periodi-ally, and having users trained and prepared appropriately& CHALLEN#ES TO IMPLEMENTIN# DR:The value of .% is not in @uestionH every organi7ation is -on-erned about its ability to get ba-6 up and running after an outage or disaster& >ut implementing .% -an be e:pensive and -omple:, as well as tedious and time! -onsuming& (owever, -onsider the alternative numerous organi7ations have simply vanished be-ause they were unable to get ba-6 to full operations rapidly after a disaster&
BENEFITS OF CLOUD COMPUTING:The various benefits of -loud -omputing are*! /$ Cost saving& 3$ Maintenan-e& 4$ %eliability&
CONCLUSION:Many options and variations for .% e:ist, and this paper highlights some of the -ommon patterns, ranging from simple ba-6up and restore to fault tolerant multi!site solutions& The +S servi-es are available on!demand and we only pay for what we use& This is a 6ey advantage for .%, where signifi-ant infrastru-ture is needed @ui-6ly, but only in the event of a disaster& "n this paper also des-ribe some free tools that are used disaster re-overy and saving the -ast& These are also play a big role in .%& This paper has shown how +S ( ma7on +eb Servi-es$ provides fle:ible, -ost!effe-tive infrastru-ture solutions, enabling us to have a more effe-tive disaster re-overy plan&
.isaster %e-overy and Cloud Servi-es
12
REFERENCES:http*FFdo-s&ama7onwebservi-es&-omF +S#C3FlatestF,ettingStarted,uideF .esigning Fault!Tolerant ppli-ations in the +S Cloud whitepaper* http*FFaws&ama7on&-omFwhitepapersF
http*FFsear-hdisasterre-overy&te-htarget&-omFtutorialFCloud!disaster! re-overy!in!todays!data!storage!environments
https*FFCloud %e-overy ! 'S .isaster %e-overy Servi-es&htm http*FFwww&i-orps&-omFManagedServi-es .isaster %e-overy&htm .isaster Prevention vs&
http*FFwww&te-hrepubli-&-omFblogFfive!appsFfive!free!disaster!re-overy! toolsF
.isaster %e-overy and Cloud Servi-es
13
Submitted To Lovely Professional University, Phagwara (For the partial fulfillment of the degree of MC !""" S#M$
SUPERVISED BY SUBMITTED BY:
M%& %'("T '(%" ( ssistant Le-turer$
.isaster %e-overy and Cloud Servi-es
)ame * P + ) S"),( %oll )o& * %./#01 21
1
%eg& )o& * //32004/
.ate of Submission* 35 '-t 32/4
ACKNOWLEDGEMENT
+ith profound regard " gratefully a-6nowledge our respe-ted -ourse Le-turer, MR. ROHIT OHRI, .epartment of C , Fa-ulty of -omputer organi7ation and ar-hite-ture, for his generous help and day today suggestion during the preparation of the report& " li6e to give than6s espe-ially to our friends 8 many individuals, for their enthusiasti- en-ouragements and helps during the preparation of this report and for their assistan-e in typing and proofreading this manus-ript&
.isaster %e-overy and Cloud Servi-es
2
Table of Content Contents Page No.
/& -6nowledgement9999999999999999993
3& "nde:99999999999999999999999&4 4& bstra-t99999999999999999999990
0& "ntrodu-tion99999999999999999999&0!; ;& Free 'f Cast Five Tools For .isaster %e-overy99999&&&<!5 <& Servi-es nd Features #ssential For .isaster %e-overy99&&5!= 1& .atabases999999999999999999999= 5& The Cloud "s "deal For .isaster %e-overy&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&=!/2 =& .isaster %e-overy "n The Cloud 'ptions99999999&/2!// /2& >enefits of -loud -omputing9999999999999&// //& Con-lusion99999999999999999999&/3 /3& %eferen-es99999999999999999999&&/3
.isaster %e-overy and Cloud Servi-es
3
ABSTRACT:+ith the growing value of data as a strategi- -orporate asset, today?s "T organi7ations fa-e the -hallenge of implementing reliable ba-6up and re-overy solutions in the most effi-ient, -ost!effe-tive manner& To meet this -hallenge, they need to -arefully define their business re@uirements and re-overy obAe-tives before de-iding on the right ba-6up and re-overy te-hnologies to deploy& "n the event of a disaster, we -an @ui-6ly laun-h resour-es in ma7on +eb Servi-es ( +S$ to ensure business -ontinuity& The paper highlights relevant +S features and servi-es that we -an leverage for our .% pro-esses and shows e:ample s-enarios about how to re-over from disaster& "t further provides re-ommendations about how we -an improve our .% plan and leverage the full potential of +S for our .isaster %e-overy pro-esses&
INTRODUCTION:DISASTER RECOVERY .isaster re-overy, 'perational ba-6up and ar-hival is used in ba-6up purpose& .isaster re-overy is a pro-ess, poli-es and pro-edure that are used to re-over the lost data& Many business and government servi-es utili7e .isaster %e-overy (.%$ systems to minimi7e the downtime raised by -atastrophisystem failures& Currently .isaster %e-overy system use periodi- tape for ba-6up& .isaster re-overy is -lassified into two -ategories, one is natural li6e storms, earth@ua6es et-& and other is made by man li6e ha7ardous materials spills, infrastru-ture failure et-& +hen any disaster re-overy is re@uired, the disaster re-overy planner first report to their organi7ation business -ontinuity plans whi-h should indi-ate the matri-es of %P' (re-overy point obAe-tive$ and %T' (re-overy time obAe-tive$&
.isaster %e-overy and Cloud Servi-es
4
THE NEED FOR BETTER DATA PROTECTION s today?s businesses in-reasingly re-ogni7e -orporate data as a strategiasset that must be prote-ted against a wide range of ris6s and threats, data prote-tion has be-ome a high!priority obAe-tive for "T organi7ations& "n turn, these organi7ations now fa-e the -hallenges of identifying, deploying, and effi-iently managing their data ba-6up and re-overy infrastru-tures& RECOVERY POINT OBJECTIVE AND RECOVERY TIME OBJECTIVE RPO* ! This des-ribes the a--eptable amount of data loss measured in time& For e:ample, if the %P' was / hour, after the system was re-overed, it would -ontain all data up to a point in time that is prior to //*22 M be-ause the disaster o--urred at noon& RTO: - This is the duration of time and the servi-e level to whi-h a business pro-ess must be restored after a disaster (or disruption$ to avoid una--eptable -onse@uen-es asso-iated with a brea6 in business -ontinuity& For e:ample, if a disaster o--urs at /3*22 PM (noon$ and the %T' is 5 hours, the .% pro-ess would ensure re-overy to the a--eptable servi-e level would be possible by 5*22 PM& CLOUD SERVICES >efore des-ribing servi-es of -loud -omputing first we 6now about -loud -omputing& The -loud -omputing is internet based -omputing, where shared all hardware and software resour-es and information to the -omputer (that are -onne-ted to internet$ on demand but paying some -harge& The various servi-es are* B /$ Software as servi-es& 3$ Platform as servi-es& 4$ "nfrastru-ture as servi-es& To minimi7e the impa-t of a disaster on the business, -ompanies invest time and resour-es to plan, prepare, do-ument, train, and update pro-esses to deal with events& The amount of investment for the disaster re-overy planning of a parti-ular system -an vary dramati-ally depending on the -ost of a potential
5
.isaster %e-overy and Cloud Servi-es
outage& This paper des-ribes some typi-al approa-hes ranging from minimal investments to full!s-ale availability and fault toleran-e& Proper preparation for .% is a must, and this paper outlines some of the best pra-ti-es to improve our .% plans and pro-esses& AWS (AMAZON WEB SERVICE) allows us to s-ale up our infrastru-ture on an as!needed basis& +e get a--ess to the same highly s-alable, reliable, se-ure, fast, ine:pensive infrastru-ture that ma7on uses to run its own global networ6 of web sites and only pay for what we use&
FREE OF COST FIVE TOOL FOR DISASTER RECOVERY:The five tools that are for disaster re-overy that are following*! /& Ma-rium %efle-ts& 3& Clone7illa& 4& .rive"mageCML& 0& Dui-6 .isaster %e-overy& ;& System %es-ue C.& 1. MACRIUM REFLECT This is the free version of the Professional produ-t& This version supports CP, Eista, +indows 1, +indows 5 and this is use for home!use des6top& For using this -an easily handle .is6 imaging, s-hedule ba-6up and also supports % ".&
Fig- .1!. !. CLONEZILLA Clone7illa is a free, open sour-e, bare metal ba-6up and re-overy tool& Clone7illa is based on .%>L, Part-lone, and Up-ast& There are two versions of Clone7illa! Live and Clone7illa S# (Server #dition$& The Live version is to be
.isaster %e-overy and Cloud Servi-es
6
used for a single des6top, whereas the Server edition is suitable for massive deployment (up to 02 ma-hine simultaneously$&
Fig- .! ". DRIVEIMA#E$ML .rive"mageCML is similar to Ma-rium %efle-t in that it offers a free version for personal use& This free version allows us to ba-6up, browse, and restores images& +ith the ability to browse images, this means we -an re-over files andFor folder (not for entire image$&
Fig- ."
%.
&UIC' DISASTER RECOVERY
Dui-6 .isaster %e-overy is a tool that -an @ui-6ly re-over a ma-hine when various built!in +indows administrator tools have been disabled (su-h as the %egistry #ditor, Tas6 Manager, et-&$
.isaster %e-overy and Cloud Servi-es
7
Fig- .% (. SYSTEM RESCUE CD System %es-ue C. is a Linu: system res-ue dis6 that allows us to administer and repair a system after a -rash& +e -an manage partitions, re-over data, edit -onfiguration files, and we -an wor6 with both Linu: and +indows systems&
Fig- .(
SERVICES
AND
FEATURES
ESSENTIAL
FOR
DISASTER RECOVERY:-
.isaster %e-overy and Cloud Servi-es
8
"n the preparation phase of .%, it is essential to -onsider the use of servi-es and features that support data migration and durable storage be-ause they enable we to restore ba-6ed up -riti-al data to +S when disaster stri6es& ma7on +eb Servi-es ( +S$ is the first Cloud vendor that 'ra-le has partnered with to enable database ba-6up in the Cloud& RE#IONS ma7on +eb Servi-es are available in multiple %egions, so we -an -hoose the most appropriate lo-ation for our disaster re-overy site& STORA#E ma7on Simple Storage Servi-e ( ma7on S4$ provides a highly durable storage infrastru-ture designed for mission -riti-al and primary data storage& COMPUTE ma7on #lasti- Compute Cloud ( ma7on #C3$ provides resi7able -ompute -apa-ity in the -loud& +ithin minutes, we -an -reate #C3 instan-es, whi-h are virtual ma-hines over whi-h we have -omplete -ontrol& NETWOR'IN# +hen dealing with a disaster, it?s very li6ely that we will have to modify networ6 settings as we are failing over to another site& A)*+,- R,./0 ;4 is a highly available and s-alable .omain )ame System (.)S$ web servi-e& "t is designed to give developers and businesses an e:tremely reliable and -ost!effe-tive way to route end users to "nternet appli-ations& ma7on .ire-t Conne-t ma6es it easy to set up a dedi-ated networ6 -onne-tion from our premise to +S&
DATABASES:For our database needs, -onsider using these +S servi-es* A)*+,- R01*/i,-*1 D*/*2*30 S045i60 (A)*+,- RDS) ma6es it easy to set up, operate, and s-ale a relational database in the -loud& Could use ma7on
.isaster %e-overy and Cloud Servi-es
9
%.S either in the preparation phase for .% (.isaster %e-overy$ to hold our -riti-al data in a running database& A)*+,- Si)710 DB is a highly available, fle:ible, non!relational data store that offloads the wor6 of database administration& "t -an also be used in the preparation and the re-overy phase of .%&
THE CLOUD IS IDEAL FOR DISASTER RECOVERY:Disaster recovery in the cloud is a relatively new concept, and like many technology trends, there's a lot of hype and misinformation out there. Cloud infrastru-ture is gaining in popularity for organi7ations that have very real business needs that rely on information te-hnology ("T$ infrastru-ture but struggle with the -osts both -apital and operational of e:panding their data -enters& >ut before even dis-ussing the -loud, it?s worthwhile to settle on a definition& .% is an ideal use -ase for ta6ing advantage of the -loud& +hile many organi7ations remain -autious about pla-ing produ-tion servi-es in the -loud, they are often more -omfortable testing those waters for .%Gespe-ially sin-e the -loud alters the e-onomi-s of .% so radi-ally& >ut disaster re-overy in the -loud isn?t a perfe-t solution, and its short-omings and -hallenges need to be -learly understood before a firm ventures into it& Se-urity usually tops the list of -on-erns*
I3 8*/* 306.4019 /4*-3:04408 *-8 3/,408 i- /;0 61,.8< H,= *40 .3043 *./;0-/i6*/08< A40 7*33=,483 /;0 ,-19 ,7/i,- ,4 8,03 /;0 61,.8 74,5i804 ,::04 3,)0 /970 ,: /=,-:*6/,4 *./;0-/i6*/i,-< D,03 /;0 61,.8 74,5i804 )00/ 40g.1*/,49 40>.i40)0-/3<
nd be-ause -louds are a--essed via the "nternet, bandwidth re@uirements also need to be -learly understood& There?s a ris6 of only planning for bandwidth
10
.isaster %e-overy and Cloud Servi-es
re@uirements to move data into the -loud without suffi-ient analysis of how to ma6e the data a--essible when a disaster stri6es*
• •
D, =0 ;*50 /;0 2*-8=i8/; *-8 -0/=,4? 6*7*6i/9 /, 408i406/ *11 .3043 /, /;0 61,.8< I: =0 71*- /, 403/,40 :4,) /;0 61,.8 /, ,--740)i303 i-:4*3/4.6/.40@ ;,= 1,-g =i11 /;*/ 403/,40 /*?0<
DISASTER RECOVERY IN THE CLOUD OPTIONS:MANA#ED APPLICATIONS AND MANA#ED DR n in-reasingly popular option is to put both primary produ-tion and disaster re-overy instan-es into the -loud and have both handled by a )*-*g08 3045i60 74,5i804 (MSP$& >y doing this we?re reaping all the benefits of -loud -omputing, from usage!based -ost to eliminating on!premises infrastru-ture& "nstead of doing it our self, we?re deferring .% to the -loud or managed servi-e provider& BAC' UP TO AND RESTORE FROM THE CLOUD ppli-ations and data remain on!premises in this approa-h, with data being ba-6ed up into the -loud and restored onto on!premises hardware when a disaster o--urs& "n other words, the 2*6?.7 i- /;0 61,.8 be-omes a substitute for tape!based off!site ba-6ups& +hen -ontemplating -loud ba-6up and re-overy, it?s -ru-ial to -learly understand both the ba-6up and the more problemati- restore aspe-ts& >a-6ing up into the -loud is relatively straightforward, and ba-6up appli-ation vendors have been e:tending their ba-6up suites with options to dire-tly ba-6 up to 7,7.1*4 61,.8 3045i60 74,5i8043 3.6; *3 ATAT@ A)*+,-@ Mi64,3,:/ C,47,4*/i,-3. The -hallenging aspe-t of using -loud!based ba-6ups for disaster re-overy is the re-overy& +ith bandwidth limited and possibly terabytes of data to be re-overed, getting data restored ba-6 on!premises within defined %T's -an be -hallenging&
.isaster %e-overy and Cloud Servi-es
11
NEW OPTIONS@ OLD FUNDAMENTALS The -loud greatly e:tends disaster re-overy options, yields signifi-ant -ost savings& "t does not, however, -hange the .% fundamentals of having to devise a solid disaster re-overy plan, testing it periodi-ally, and having users trained and prepared appropriately& CHALLEN#ES TO IMPLEMENTIN# DR:The value of .% is not in @uestionH every organi7ation is -on-erned about its ability to get ba-6 up and running after an outage or disaster& >ut implementing .% -an be e:pensive and -omple:, as well as tedious and time! -onsuming& (owever, -onsider the alternative numerous organi7ations have simply vanished be-ause they were unable to get ba-6 to full operations rapidly after a disaster&
BENEFITS OF CLOUD COMPUTING:The various benefits of -loud -omputing are*! /$ Cost saving& 3$ Maintenan-e& 4$ %eliability&
CONCLUSION:Many options and variations for .% e:ist, and this paper highlights some of the -ommon patterns, ranging from simple ba-6up and restore to fault tolerant multi!site solutions& The +S servi-es are available on!demand and we only pay for what we use& This is a 6ey advantage for .%, where signifi-ant infrastru-ture is needed @ui-6ly, but only in the event of a disaster& "n this paper also des-ribe some free tools that are used disaster re-overy and saving the -ast& These are also play a big role in .%& This paper has shown how +S ( ma7on +eb Servi-es$ provides fle:ible, -ost!effe-tive infrastru-ture solutions, enabling us to have a more effe-tive disaster re-overy plan&
.isaster %e-overy and Cloud Servi-es
12
REFERENCES:http*FFdo-s&ama7onwebservi-es&-omF +S#C3FlatestF,ettingStarted,uideF .esigning Fault!Tolerant ppli-ations in the +S Cloud whitepaper* http*FFaws&ama7on&-omFwhitepapersF
http*FFsear-hdisasterre-overy&te-htarget&-omFtutorialFCloud!disaster! re-overy!in!todays!data!storage!environments
https*FFCloud %e-overy ! 'S .isaster %e-overy Servi-es&htm http*FFwww&i-orps&-omFManagedServi-es .isaster %e-overy&htm .isaster Prevention vs&
http*FFwww&te-hrepubli-&-omFblogFfive!appsFfive!free!disaster!re-overy! toolsF
.isaster %e-overy and Cloud Servi-es
13
