Bigtable: A Distributed Storage System for Structured Data : part11 Conclusions and Acknowledgements

11 Conclusions
We have described Bigtable, a distributed system for storing structured data at Google. 
Bigtable clusters have been in production use since April 2005, and we spent roughly seven person-years on design and implementation before that date. 
As of August 2006, more than sixty projects are using Bigtable. 
Our users like the performance and high availability provided by the Bigtable implementation, and that they can scale the capacity of their clusters by simply adding more machines to the system as their resource demands change over time.


11结论
我们已经描述了Bigtable，一种用于在Google存储结构化数据的分布式系统。
自2005年4月以来，Bigtable集群已投入使用，在此之前，我们花费了大约7个人的时间来设计和实施。
截至2006年8月，60多个项目正在使用Bigtable。
我们的用户喜欢Bigtable实现提供的性能和高可用性，并且通过在资源需求随时间变化的情况下，通过简单地向系统添加更多的机器来扩展其集群的容量。


Given the unusual interface to Bigtable, an interesting question is how difficult it has been for our users to adapt to using it. 
New users are sometimes uncertain of how to best use the Bigtable interface, particularly if they are accustomed to using relational databases that support general-purpose transactions. 
Nevertheless, the fact that many Google products successfully use Bigtable demonstrates that our design works well in practice.
We are in the process of implementing several additional Bigtable features, such as support for secondary indices and infrastructure for building cross-data-center replicated Bigtables with multiple master replicas. 
We have also begun deploying Bigtable as a service to product groups, so that individual groups do not need to maintain their own clusters. 
As our service clusters scale,we will need to deal with more resource-sharing issues within Bigtable itself.
Finally, we have found that there are significant advantages to building our own storage solution at Google.
We have gotten a substantial amount of flexibility from designing our own data model for Bigtable. 
In addition, our control over Bigtable’s implementation, and the other Google infrastructure upon which Bigtable depends, means that we can remove bottlenecks and inefficiencies as they arise.


鉴于BigTable的不寻常接口，一个有趣的问题是我们的用户适应使用它有多困难。
新用户有时不确定如何最好地使用Bigtable界面，特别是如果他们习惯于使用支持通用交易的关系数据库。
然而，许多Google产品成功使用Bigtable的事实表明我们的设计在实践中表现良好。
我们正在实施几个额外的BigTable功能，例如支持辅助索引和基础架构，用于构建具有多个主副本的跨数据中心复制的Bigtables。
我们也已经开始将Bigtable作为服务部署到产品组，以便个别组织不需要维护自己的集群。
随着我们的服务集群规模的扩大，我们将需要处理Bigtable中更多的资源共享问题。
最后，我们发现在Google上构建自己的存储解决方案有很大的优势。
我们从设计我们自己的Bigtable数据模型中获得了很大的灵活性。
此外，我们对BigTable的实施以及Bigtable所依赖的其他Google基础架构的控制意味着我们可以消除瓶颈和低效率。




Acknowledgements
We thank the anonymous reviewers, David Nagle, and our shepherd Brad Calder, for their feedback on this paper. 
The Bigtable system has benefited greatly from the feedback of our many users within Google. 
In addition,we thank the following people for their contributions to Bigtable: Dan Aguayo, Sameer Ajmani, Zhifeng Chen, Bill Coughran, Mike Epstein, Healfdene Goguen, Robert Griesemer, Jeremy Hylton, Josh Hyman, Alex Khesin, Joanna Kulik, Alberto Lerner, Sherry Listgarten, Mike Maloney, Eduardo Pinheiro, Kathy Polizzi, Frank Yellin, and Arthur Zwiegincew.


致谢
我们感谢匿名评审人David Nagle和我们的牧羊人Brad Calder对本文的反馈意见。
Bigtable系统受益于Google内许多用户的反馈。
此外，我们感谢以下人士对Bigtable的贡献：Dan Aguayo，Sameer Ajmani，陈志峰，Bill Coughran，Mike Epstein，Healfdene Goguen，Robert Griesemer，Jeremy Hylton，Josh Hyman，Alex Khesin，Joanna Kulik，Alberto Lerner， Sherry Listgarten，Mike Maloney，Eduardo Pinheiro，Kathy Polizzi，Frank Yellin和Arthur Zwiegincew。