Final Notes on Architecture (Large-Scale)

even more!! aaaaaaa

what can we do?

think big, manage the complexity of structure as it grows, concentrate on the big picture

location transparency

interact with any service regardless of location

failure transparency

if a machine goes down, the service continues uninterrupted

load transparency

no matter how many people use the system, it stays responsive

upgrade transparency

a component can be changed, moved, maintained or added to without bringing the system down

what do we use http or similar transport protocols for

fetching data back to the local site, and sometimes write data back

what do we use remote object protocols for

make the remote site perform some action

how do we identify things?

universal resource identifiers (URI), interoperable object reference (IOR) in CORBA

what 2 issues are there with locating?

how does the client find the database? and how do all clients find the same database?

how do we handle the database location issue?

don’t hard-wire it because it will change. instead, look it up online in a directory so there is a single point of change

how does replication work?

client can use the first database but if it’s too slow it can switch to the second one. typically put the replicas on different machines/ networks/ sites/ continents

what two issues do we face with replication?

handling local copies and load balancing