2010-11-24

Hadoop & GPFS

• ​IBM Builds on Hadoop with New Storage Architecture - IBM 讓 GPFS 可以跟 HDFS 合作，提供更高可用度的儲存服務。稱之為『General Parallel File System-Shared Nothing Cluster (GPFS-SNC)』
• ​IBM發表新儲存架構 GPFS-SNC

Public Large Data Set 公用大型資料集

• 維基百科 - ​http://en.wikipedia.org/wiki/Wikipedia:Database_download
• 亞馬遜提供的公用資料 - ​http://aws.amazon.com/publicdatasets/
  • 包括基因資料(Genome, Ex. 1000 Genome Project)
• ​http://www.statmt.org/europarl/
• ​http://www.opendatacenteralliance.org/
• ​Data.gov - 美國公部門的資料 - 台灣應該可以上 ​檔案管理局 (2010-11-08)
• ​http://ckan.org/
• ​http://stat-computing.org/dataexpo/2009/ - 飛航紀錄 - 1987-2008 years of airlines performance data. 1GB, 12M records.
• ​http://kdd.ics.uci.edu/ - UCI Knowledge Discovery in Databases Archive - The data sets are categorized neatly. These are very useful for many machine-learning exercises with the use of Hadoop/Mahout?.
• A list of about 70 data sets compiled for Open Data Day: ​http://www.opendataday.org/wiki/Data
• A list of 70 datasets from technology review: ​https://www.technologyreview.com/blog/arxiv/26097/
• Stanford large network database collection: ​http://snap.stanford.edu/data/index.html
• Digging into data dataset list: ​http://www.diggingintodata.org/Repositories/tabid/167/Default.aspx
• Many Eyes from IBM: ​http://www-958.ibm.com/software/data/cognos/manyeyes/
• Data360: ​http://www.data360.org/index.aspx
• UN Data Explorer: ​http://data.un.org/Explorer.aspx
• OECD Data Sets: ​http://stats.oecd.org/index.aspx
• Freebase data dumps: ​http://wiki.freebase.com/wiki/Data_dumps
• ​http://infochimps.org - They have over a billion tweets.
• ​http://snap.stanford.edu/data/ - 史丹佛大學的一些資料集
• ​http://data.stackexchange.com/