文本分析教案課程anti cache_第1頁
文本分析教案課程anti cache_第2頁
文本分析教案課程anti cache_第3頁
文本分析教案課程anti cache_第4頁
文本分析教案課程anti cache_第5頁
已閱讀5頁,還剩15頁未讀, 繼續(xù)免費閱讀

下載本文檔

版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進(jìn)行舉報或認(rèn)領(lǐng)

文檔簡介

1、The cache mechanism in H-Storeassumption: data wont fit in memory disk-resident data, main memory buffer pool for execution concurrency is a must transaction serialization and locksDBMS buffer pooldistributed cachein-memory DBMSmust still maintain buffer pool lock/latch data maintain ARIES-style rec

2、overy logsquestion: What is the overhead of all these things?fast and scalable, butkey-value interfacenot ACID (AI, not CD)H-Store Architecture partitioned, shared-nothing single-threaded main memory execution no need for locks and latcheslightweight recovery snapshots + command logbig: disk-oriente

3、dfast: memory-orientedbig and fast: anti-cachingasynchronous disk fetches dont blockmaintain ordering of evicted data accesses ensures transactional consistencysingle copy of data consistency is freeefficient memory use, no swizzlingmemory is primary storage, cold data is evicted to disk-based anti-

4、cachereading data from the anti-cache is done in 3 phases avoids blocking, ensures consistencyBlock Table This is a hash table that maintains the blocks of tuples that have been evicted from the DBMSs main memory storage.Evicted Table The Evicted Table keeps track of the tuples that have been writte

5、n out to blocks on disk. DBMS updates any indexes containing evicted tuples to reference the Evicted Table.LRU Chain: Lastly, H-Store also maintains an in-memory list of all the tuples for each table in LRU order.Evict 1. data anti-cache threshold 2. dynamically construct anticache blocks of coldest

6、 tuples 3. asynchronously write to diskpre-pass 1. a transaction enters pre-pass when evicted data is accessed 2. continues execution, creating list of evicted blocks 3. abort, queue blocks to be fetchedFetch 1. data is fetched asynchronously from disk and avoids blocking 2. moved into merge bufferM

7、erge 1. data is moved from in-memory merge buffer to in-memory table. Block-merging or Tuples-merging 2. previously aborted transaction is restarted 3. transaction executes normallydone online, more responsive to changes in workloadgoal is low CPU and memory overheadapproximate ordering is OKmaintain LRU chain embedded in tuple headersper-partitiontransactions that update LRU chain are sampled randomly configurable sample ratefine-grained evictionblocks constructed dynamicallyasynchronous batched fetchespossible because of transactionsdata exists in exactly one location caching archi

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
  • 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負(fù)責(zé)。
  • 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

最新文檔

評論

0/150

提交評論