RECIPE : Converting Concurrent DRAM Indexes to Persistent-Memory Indexes (SOSP 2019)

RECIPE proposes a principled approach for converting concurrent indexes built for DRAM into crash-consistent indexes for persistent memory. This repository includes the implementations of the index structures for persistent memory converted from the existing concurrent DRAM indexes by following RECIPE. For performance evaluations, this repository also provides the microbenchmarks for index structures based on YCSB. This repository contains all the information needed to reproduce the main results from our paper.

Please cite the following paper if you use the RECIPE approach or RECIPE-converted indexes:

RECIPE : Converting Concurrent DRAM Indexes to Persistent-Memory Indexes. Se Kwon Lee, Jayashree Mohan, Sanidhya Kashyap, Taesoo Kim, Vijay Chidambaram. Proceedings of the The 27th ACM Symposium on Operating Systems Principles (SOSP 19). Paper PDF. Extended version(arXiv). Bibtex.

Integrating RECIPE indexes into your own project

Apart from benchmark code with ycsb.cpp, we provide simple example codes (P-*/example.cpp for each RECIPE index) to help developers who want to apply RECIPE indexes into their own project to easily identify how to use each index's APIs. These example source codes run insert and lookup operations with custom integer keys. For more details of usage for each index, please refer to P-*/README.md in each index's directory and ycsb.cpp as well.

Contents

1. P-CLHT/ contains the source code for P-CLHT. It is converted from Cache-Line Hash Table to be persistent. The original source code and paper can be found in code and paper.
2. P-HOT/ contains the source code for P-HOT. It is converted from Height Optimized Trie to be persistent. The original source code and paper can be found in code and paper.
3. P-BwTree/ contains the source code for P-BwTree. It is converted from an open sourced implementation of BwTree for persistent memory. The original source code and paper can be found in code and paper.
4. P-ART/ contains the source code for P-ART. It is converted for persistent memory from Adaptive Radix Tree using ROWEX for concurrency. The original source code and paper can be found in code and paper.
5. P-Masstree/ contains the source code for P-Masstree. It is converted from Masstree to be persistent and is custumized for the compact version. The original source code and paper can be found in code and paper.
6. index-microbench/ contains the benchmark framework to generate YCSB workloads. The original source code can be found in code.

Recommended use cases for RECIPE indexes

1. P-CLHT is a good fit for applications requiring high-performance point queries.
2. P-HOT is a good fit for applications with read-dominated workloads.
3. P-BwTree provides well-balanced performance for insertion, lookup, and range scan operations for applications using integer keys.
4. P-ART is suitable for applications with insertion-dominated workloads and a small number of range queries.
5. P-Masstree provides well-balanced performance for insertion, lookup, and range scan operations for applications using either integer or string keys.

Running RECIPE Indexes on Persistent Memory and DRAM

Desired system configurations (for DRAM environment)

• Ubuntu 18.04.1 LTS
• At least 32GB DRAM
• x86-64 CPU supporting at least 16 threads
• P-HOT: x86-64 CPU supporting at least the AVX-2 and BMI-2 instruction sets (Haswell and newer)
• Compile: cmake, g++-7, gcc-7, c++17

Dependencies

Install build packages

$ sudo apt-get install build-essential cmake libboost-all-dev libpapi-dev default-jdk

Install jemalloc and tbb

$ sudo apt-get install libtbb-dev libjemalloc-dev

Generating YCSB workloads

Download YCSB source code

$ cd ./index-microbench
$ curl -O --location https://github.com/brianfrankcooper/YCSB/releases/download/0.11.0/ycsb-0.11.0.tar.gz
$ tar xfvz ycsb-0.11.0.tar.gz
$ mv ycsb-0.11.0 YCSB

How to configure and generate workloads

Configure the options of each workloads (a, b, c, e), would only need to change $recordcount and $operationcount.

$ vi ./index-microbench/workload_spec/<workloada or workloadb or workloadc or workloade>

Select which workloads to be generated. Default configuration will generate all workloads (a, b, c, e). Change the code line for WORKLOAD_TYPE in <a b c e>; do, depending on which workload you want to generate.

$ vi ./index-microbench/generate_all_workloads.sh

Generate the workloads. This will generate both random integer keys and string ycsb keys with the specified key distribution.

$ cd ./index-microbench/
$ mkdir workloads
$ bash generate_all_workloads.sh

Checklists

Configuration for workload size.

Change LOAD_SIZE and RUN_SIZE variables to be same with the generated workload size, which are hard-coded in ycsb.cpp (Default is 64000000).

$ vi ycsb.cpp

Configurations for Persistent Memory

For running the indexes on Intel Optane DC Persistent Memory, we will use libvmmalloc to transparently converts all dynamic memory allocations into Persistent Memory allocations, mapped by pmem.

Ext4-DAX mount

$ sudo mkfs.ext4 -b 4096 -E stride=512 -F /dev/pmem0
$ sudo mount -o dax /dev/pmem0 /mnt/pmem

Install PMDK

$ git clone https://github.com/pmem/pmdk.git
$ cd pmdk
$ git checkout tags/1.6
$ make -j
$ cd ..

Configuration for libvmmalloc

• LD_PRELOAD=path

Specifies a path to libvmmalloc.so.1. The default indicates the path to libvmmalloc.so.1 that is built from the instructions installing PMDK above.

• VMMALLOC_POOR_DIR=path

Specifies a path to the directory where the memory pool file should be created. The directory must exist and be writable.

• VMMALLOC_POOL_SIZE=len

Defines the desired size (in bytes) of the memory pool file.

$ vi ./scripts/set_vmmalloc.sh

Please change below configurations to fit for your environment.

export VMMALLOC_POOL_SIZE=$((64*1024*1024*1024))
export VMMALLOC_POOL_DIR="/mnt/pmem"

Building & Running on Persistent Memory and DRAM

Build all

$ mkdir build
$ cd build
$ cmake ..
$ make

DRAM environment

Run

$ cd ${project root directory}
$ ./build/ycsb art a randint uniform 4
Usage: ./ycsb [index type] [ycsb workload type] [key distribution] [access pattern] [number of threads]
       1. index type: art hot bwtree masstree clht
                      fastfair levelhash cceh
       2. ycsb workload type: a, b, c, e
       3. key distribution: randint, string
       4. access pattern: uniform, zipfian
       5. number of threads (integer)

Persistent Memory environment

Run

$ cd ${project root directory}
$ sudo su
# source ./scripts/set_vmmalloc.sh
# LD_PRELOAD="./pmdk/src/nondebug/libvmmalloc.so.1" ./build/ycsb art a randint uniform 4
# source ./scripts/unset_vmmalloc.sh

PMDK Implementation

The RECIPE indices are currently being ported to PMDK to ensure the recoverability of each data structure. Since RECIPE does not handle memory leaks during crashes, PMDK ensures the user's ability to recover any allocations that were made in a persistent memory pool. Currently, only the CLHT data structure has been ported to use PMDK. For more details regarding the PMDK implementation please check out the pmdk branch as well as these details.

Artifact Evaluation

For artifact evaluation, we will evaluate again the performance of the index structures presented in the paper by using YCSB benchmark. The index structures tested for artifact evaluation include P-CLHT P-ART, P-HOT, P-Masstree, P-Bwtree, FAST&FAIR, WOART, CCEH, and Level hashing. The evaluation results will be stored in ./results directory as csv files. Please make sure to check the contents at least by checklists subsection in Benchmark details section below, before beginning artifact evaluation. Note that the evaluations re-generated for artifact evaluation will be based on DRAM because Optane DC persistent memory machine used for the evaluations presented in the paper has the hard access limitation from external users. For more detail, please refer to experiments.md.

RECIPE has been awarded three badges: Artifact Available, Artifact Functional, and Results Reproduced.

Limitations

1. Current implementations are based on general volatile memory allocation API such as malloc, posix_memalign, new, and etc. Just for performance testing on real PM, you can use libvmmalloc, which transparently converts all the dynamic memory allocations into Persistent Memory allocations. However, if you want to apply RECIPE indexes into your real PM application, you would need to change current volatile memory allocators using libpmem APIs. Currently, a version of RECIPE is being developed using the PMDK library. Check out the pmdk branch!

2. Current implementations only ensure the lowest level of isolation (Read Uncommitted) when using them for transactional systems, since they are based on normal CASs and temporal stores coupled with cache line flush instructions. However, you may extend them to guarantee the higher level of isolation (Read Committed) by employing alternative primitives such as either Link-and-Persist (paper, code) or PSwCAS (paper, code) and non-temporal stores coupled with memory fence.

License

The licence for most of the P-* family of persistent indexes is Apache License (https://www.apache.org/licenses/LICENSE-2.0). This is consistent with the most of the indexes we build on, with the exception of CLHT and HOT, which uses the MIT and ISC License respectively. Accordingly, P-CLHT is under the MIT license (https://opensource.org/licenses/MIT). P-HOT is under the ISC license (https://opensource.org/licenses/ISC).

Copyright for RECIPE indexes is held by the University of Texas at Austin. Please contact us if you would like to obtain a license to use RECIPE indexes in your commercial product.

Acknowledgements

We thank the National Science Foundation, VMware, Google, and Facebook for partially funding this project. We thank Intel and ETRI IITP/KEIT[2014-3-00035] for providing access to Optane DC Persistent Memory to perform our experiments.

Contact

Please contact us at [email protected] and [email protected] with any questions.