Haoyuan Li presents at Beijing Meetup on open source data orchestration and the value of leveraging Alluxio with rising trends driving the need for a new architecture. Four big trends driving this need: Separation of compute & storage, hybrid-multi cloud environments, rise of object store and self-service data across the enterprise.
This talk shares our design, implementation and optimization of Alluxio metadata service to address the scalability challenges, focusing on how to apply and combine techniques including tiered metadata storage (based on off-heap KV store RocksDB), fine-grained file system inode tree locking scheme, embedded state-replicate machine (based on RAFT), exploration and performance tuning in the correct RPC frameworks (thrift vs gRPC) and etc.
As the data ecosystem becomes massively complex and more and more disaggregated, data analysts and end users have trouble adapting and working with hybrid environments. The proliferation of compute applications along with storage mediums leads to a hybrid model that we are just not accustomed to.
With this disaggregated system data engineers now come across a multitude of problems that they must overcome in order to get meaningful insights.
Twitter SF is hosting 2019’s half yearly RocksDB Meetup with speakers from Twitter, Facebook and the community on July 11th.
Join us June 24 in Menlo Park for our next meetup! We’ll have 3 valuable talks, a delicious BBQ dinner and amazing summertime-themed raffle prizes! This free event is sponsored by GridGain Systems and Oracle.
Traditionally, if you want to run a single Spark job on EMR, you might follow the steps: launching a cluster, running the job which reads data from storage layer like S3, performing transformations within RDD/Dataframe/Dataset, finally, sending the result back to S3. You end up having something like this.
If we add more Spark jobs across multiple clusters, you could have something like this.
The latest advances in container orchestration by Kubernetes bring cost savings and flexibility to compute workloads in public or hybrid cloud environments. On the other hand, it introduces new challenges such as how to move data to compute efficiently, how to unify data across multiple or remote clouds, how to co-locate data with compute and many more. Alluxio approaches these problems in a new way. It helps elastic compute workloads realize the true benefits of the cloud, while bringing data locality and data accessibility to workloads orchestrated by Kubernetes
A new generation of open source big data, represented by Alluxio, born at the University of California at Berkeley, looks at this issue. Different from systems such as designing storage tight coupling to achieve low-cost reliable storage HDFS, by providing a virtual data storage layer defined and implemented by software for data applications, abstracting and integrating cloudy, hybrid cloud, multi-data center and other environments The underlying files and objects, and through intelligent workload analysis and data management, make data close to computing and provide data locality, big data and machine learning applications can be achieved with the same performance and lower cost.
Hear how DBS Bank is taking a new approach to making data-intensive compute independent of the storage. They will share the challenges as well as the new technology stack that includes technologies like Spark, Alluxio and object stores.