Distributed Systems PDF

STRI BUTE D DI SYST TE M S INTRODUCTION to DIST RIB U TE S T E M S D SY  DISTRIBUTED SYSTEMS is a c o ll ec tio n o f c o mp u te r A distribute d s ys t e m e c o m pu t a t ion a l re so ur c es progra...

STRI BUTE D DI SYST TE M S INTRODUCTION to DIST RIB U TE S T E M S D SY  DISTRIBUTED SYSTEMS is a c o ll ec tio n o f c o mp u te r A distribute d s ys t e m e c o m pu t a t ion a l re so ur c es programs t ha t u t il iz a r at e c o m p u ta tio n no de s t o across m ult iple , s ep m m o n, s h a re d go al. achieve a co  istributed System is Important? Why D a re im po rt a n t b ec a u s e t he y Distributed sy st e m s a n ta ge s ov e r c e nt r a liz ed offer se v er a l a d v e m w e l l- su ite d f o r a w id e systems, ma k in g t h io n s a n d u se ca s es. T h e range of a p pl ica t tr ibu t ed c o m p ut in g in c lu d e : advanta ge s o f d is  crease Wo rkloads and User 1. Adapts to the In Demands h a n dl e gr o w in g w o rk lo ad s Distributed sys te m s u se r de m a n d s b y ad di n g m o re and increasin g er v er s o r p ro c e ss in g po w er. resources, such a s s  inuity in the Face of Failures 2. Ensures System Cont co m po n e n t o r n o d e fa il s, t h e em , if o ne In a distributed syst it h o u t m ca n co n t in ue fu n ct io n ing w rest of the sy st e co rp o r a t in g r e du n da n c y a n d significant disruptio n. B y in r ib ut e d s y s t e m s m in im iz e t h e replication st ra t eg ie s , d is t o n o v e r a l l p er f o rm a nc e a n d impact of syst e m f a il u re s availability.  nce and Utilization through 3. Boosts Performa Collaboration en a bl e th e e ff icie n t sh a rin g Distributed sys t em s p ro c ess in g p o w e r, s to r a ge, of resources, such as c at io n n et w o r k ba n dw id t h. and comm un i  er ie nce w it h Geographically 4. Enhances User Ex p Dist rib ut ed Syst em s h ic a ll y d is t rib u t ed s y st e m s, In geograp te g ic a l l y l o c a t ed cl o s e r t o components are s tr a re d u ce s t h e t im e it t a k es fo r end-users, whic h e en t he u se r an d t h e s y s t em. data to travel b e tw  orm ance During High Demand 5. Maintains System Per f Periods s o f dis t r ib u tin g w o r k lo ad s Load balancin g is t h e p r o ce s o m p o n e n ts o r n o d es w it h in a evenly among m u l t ip le c r ibu t e d a r ch it ec t ur e e n s ur e s distributed sys t e m. T h is d ist n t b e c o m e s o v er l o ad e d w it h that no sing l e co m p o ne in sy s t e m p er f o r m an ce ev e n requests, helpin g t o m ain t a during high demand. DISTRIBUTED Computi ng Mo d e l s Physical Model - o w d e v ices w a r e co m p o n e n t s a n d h Represents h a r d d ist r ib u te d s y s t em s. interconnect in Key Components: u n ic a t e. th a t p r oc e s s d a t a a n d c o m m Nodes: En d d e v ice s a n n e ls (w ir ed /w ir e le s s ). Links: Communication ch e f or de c e n t r a l iz ed c o nt r o l. Midd le w a re : So ft w a r f n o de s (b u s , s t a r, m e sh , e t c.). Topolog y : A r r a n g e m e n t o m u n ic a t io n (T C P , U D P , H T T PS ). Proto c ol s : R u le s fo r c o m. Architectural model - iz a tio n o f a sig n, st r u ct u r e, a n d o r g a n Describes t he de distributed system. Key components: r eq ue s ts a nd t r a liz e d s y st e m w it h cl ie n t Client-Server: C en server responses. od e s c a n D e c en t ra l iz ed , w he r e a l l n Peer-to-P ee r (P 2P ): reques t /pr o vid e s er v ic es. r s , ea c h p ro vid in g a s pec if ic r g an iz e d in t o la ye Layered: O service. Fundamental model - a m e w o r k f o r un d e r s t a n d in g Co n c ep tu a l f r t ed s y s te m b e h a v io r. distribu key components: t e e l : H o w p r o c es s es c o m mu n ic a 1. Intera c tio n M o d s in g , P u b lis h /S u b s cr ib e , R P C ). (M es s a g e Pa s ib es sy s te m f a il u re s ( C r a sh , 2. Failure M o d el : D es c r Omission, T im in g , B y za n t in e ). cu se s o n p r o t e ct io n a g a in st 3. Securit y M o d e l : F o s a n d u n a u t ho r iz e d a c c es s. ma l ic io u s a t t a c k Ben ef its of Dist rib u te d Co m p u tin g  fits of Distributed Computing Bene - Improved P e rf o rm an ce - Enhanced D at a Pr o ce ss in g b il it y an d F au lt - Greater Relia Tolerance  H AL LE NG E S IN C TRI BU TE D SY S T EMS DIS nce Fault Tolera Consistency and  o f fe r g r ea t a d va n ta g es lik e Distributed sy st e m s a rin g, a nd hig h a v ail ab ilit y. scalability, resou rc e s h fit s c o m e u niq ue c ha l le ng es. However, with t h es e b en e a tion , d at a c on siste nc y , an d Managing com m u nic inc re as ing ly co mp le x a s the handling failures be co me system scales.  Consistency ensures that all nodes in a distributed system reflect the same data at any given time. 1. Consistency However, maintaining strict consistency across the network can be challenging, especially when considering the CAP Theorem. What is CAP Theorem? According to the CAP Theorem, a distributed system can only guarantee two out of three properties: Consistency, Availability, and Partition Tolerance. This means trade-offs are often necessary, especially when network issues arise. To address this, many systems adopt a model called eventual consistency eventual consistency Here, while temporary inconsistencies are allowed, all nodes will eventually converge to a consistent state. This approach maintains availability and performance while acknowledging the complexity of ensuring strict, immediate consistency.  Fault tolerance is a system’s ability to continue operating even when some components fail. This is achieved through redundancy and replication, which involve storing multiple copies of data across 2. Fault Tolerance nodes. If one node fails, other copies ensure that data remains accessible, preserving the system’s reliability. how do manage these copies and keep them synchronized? consensus algorithms To manage these copies and keep them synchronized, distributed systems rely on consensus algorithms like Paxos and Raft. These algorithms help nodes agree on a shared state, ensuring that the system can function correctly despite failures or inconsistencies. Together, redundancy and consensus enable the system to be both resilient and dependable. THhA a N nkK Y OU !