Data Loading Performance

Typical Issues:

• Loads are too slow to keep up with analysis needs
• Loading limited to nightly batch windows

Causes:

• Inability of loading to scale
  • One node loading serially, instead of in parallel
• Co-dependent processing
  • Queries and loading on same nodes
  • Query performance degrades or causes downtime
  • Your business cannot tolerate a query outage, delaying loading
• Destination bottlenecks
  • One node receiving loads serially
  • Data is queued up and bottlenecked by disk write speed of single node
• Network becomes a bottleneck
  • Too much data being loaded through a congested interconnect

Remedies:

• Always-parallel performance and scalability
  • On-demand incremental scaling of loading capacity
  • Meet any loading service level agreement
• High-volume loading/exporting
  • Loading and exporting scales linearly with additional nodes
  • Loading/exports/backup is concurrent with query processing
• Massively-parallel, network-aware database architecture
  • Leverage deep awareness of network topology