MagistraleOS

Argomenti

12 argomenti · padronanza manuale

What is the fundamental difference between vertical and horizontal scaling?
Why does vertical scaling eventually hit a hard physical ceiling?
Why do big data systems prefer many commodity machines over few premium ones?
Why must failure be treated as a "normal" event in scale-out systems, rather than an exception?
Why do big data DFSs use large block sizes instead of small ones?
Why is metadata kept separate from the actual data storage servers?
Walk through the client interaction model: what does the metadata server not do?
Why is RAID-style protection insufficient at cluster scale?
What two types of information does the GFS Master manage, and why is metadata kept entirely in memory?
Walk through the GFS read path: what is the Master's role, and what is the chunkserver's role?
Explain the function of the chunk lease mechanism in the GFS write path.
How does GFS tolerate a Master crash without losing metadata?

● Rosso: non ne ho idea ● Giallo: lo so, ma non benissimo ● Verde: lo so perfettamente

0 lezioni registrate

Registra la prima lezione con i tuoi appunti.

0 di 0 completate

Aggiungi la prima attività per questo esame.

Elimina esame

Verranno eliminati anche argomenti, diario, attività e ore di studio.

Ogni minuto investito nella tua preparazione conta.

Ore messe da parte

2 h 0 min

2 versamenti di studio

Storico dei versamenti (2)