Administration

Typical big data analytics management issues:

• Unscheduled system downtime
• Tedious management of existing system
• Constantly increasing temporary storage space and/or memory

Causes:

• Scaling is hard
  • Designing and setting up new hardware takes too long
  • Configuring new hardware to work with older generations is error-prone or impossible
  • Common tasks for installing new nodes are manual: repartitioning, re-indexing, or re-creating constraints
• System is not designed for fault-tolerance
  • Integrating high availability and recovering from failures is hard
  • Restoring the replicas/backups when hardware components fail is complex and time-consuming
  • Detecting and recovering from transient software failures is difficult or impossible
  • System management is difficult
  • No single-system view of the data and analytics workloads
  • Queries cannot be monitored
  • Run-away queries cannot be cancelled
  • Adding/removing hardware to the cluster is time consuming and error-prone
  • Constantly tuning the system is time consuming and difficult

Remedies:

• Always-parallel performance and scaling
  • Incremental scalability
  • Heterogeneous commodity scale-out
• Always-on availability and resiliency
  • Live system administration
  • One-click scaling
  • Online fault-tolerance and recovery
  • Online replica restoration
  • Online back-up/restore
• Cutting-edge manageability