bk99.de entertain the web since 1997

bk99.de
HN · ArchivingCurated

Amazon Glacier for rarely needed long-term data

Summary

Amazon Glacier introduces inexpensive object storage whose delayed retrieval is aimed at archives rather than interactive use. AWS announced Glacier in August 2012. The service was aimed at rarely read data.

Ideas

  • Low storage costs are traded for longer recovery times.
  • Archives need different access patterns from active data.
  • Retrieval costs depend on volume, speed and number of requests.
  • Lifecycle rules automatically move old objects into archive classes.

Insights

  • Storage classes turn time into a calculable resource.
  • An archive is only valuable once search and recovery work.
  • Cheap storage can become expensive through poorly planned retrievals.

Facts

  • Retrievals could take several hours.

Recommendations

  • Regularly test restores and the time limits you need.
  • Document keys, formats and retention rules outside the archive.

References

Read the original article

Search the Web Archive

HN · 2008

Amazon EBS separates persistent data from EC2 instances

Amazon introduces Elastic Block Store, giving short-lived EC2 instances independently persisting block devices with S3 snapshots. Amazon announced EBS in August 2008. EBS provided persistent block storage for EC2.

HN · 2013

Data structures and algorithms in the Linux kernel

A curated overview shows where the Linux kernel uses trees, hash lists, heaps, compression and scheduling algorithms in practice. The kernel provides generic list and tree helpers. CFS uses a red-black tree.

HN · 2013

Big O as a reference for data structures

Eric Rowell compares the running time and memory requirements of common data structures and sorting algorithms in a compact overview. The overview contains search, insertion and deletion costs. Several sorting algorithms are compared graphically.