Using vector interfaces to deliver millions of IOPS from a networked key-value storage server

Vasudevan, Vijay; Kaminsky, Michael; Andersen, David G.

doi:10.1145/2391229.2391237

Cited by 24 publications

(12 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[17] presents M-Lock, which is a framework that helps accelerate distributed transactions on key-value stores in cloud computing platforms. Finally, [18] attempts to address the increasing software-I/O gap in key-value stores by using vector interfaces in high-performance networked systems as the basis for key-value storage servers, where they demonstrate that they can provide 1.6 million requests per second with a median latency below one millisecond thanks to the high speed of non-volatile memories. All these previous efforts are different from ours in the sense that they are targeting at accurate key-value stores while our approach is approximate by nature.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

kBF: Towards Approximate and Bloom Filter based Key-Value Storage for Cloud Computing Systems

Xiong

Yao

et al. 2017

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

As one of the most popular cloud services, data storage has attracted great attention in recent research efforts. Key-value (k-v) stores have emerged as a popular option for storing and querying billions of key-value pairs. So far, existing methods have been deterministic. Providing such accuracy, however, comes at the cost of memory and CPU time. In contrast, we present an approximate k-v storage for cloud-based systems that is more compact than existing methods. The tradeoff is that it may, theoretically, return errors. Its design is based on the probabilistic data structure called "bloom filter", where we extend the classical bloom filter to support key-value operations. We call the resulting design as the kBF (key-value bloom filter). We further develop a distributed version of the kBF (d-kBF) for the unique requirements of cloud computing platforms, where multiple servers cooperate to handle a large volume of queries in a load-balancing manner. Finally, we apply the kBF to a practical problem of implementing a state machine to demonstrate how the kBF can be used as a building block for more complicated software infrastructures.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Among its various forms, key-value (k-v) stores have emerged as a popular option for storing and querying billions of key-value pairs [1], [2], [3] [4]. Examples of such services include Amazon Dynamo [5], Memcached [6] (used by Facebook and Twitter, etc.…”

Section: Introductionmentioning

confidence: 99%

kBF: Towards Approximate and Bloom Filter based Key-Value Storage for Cloud Computing Systems

Xiong

Yao

et al. 2017

IEEE Trans. Cloud Comput.

View full text Add to dashboard Cite

show abstract

“…Since we only know their XOR result, and the encoding scheme we designed earlier does not guarantee that the encodings give unique values when three or more of them are combined, we only provide an opportunistic approach for three-item decoding. For more than three items that are mapped to the same cell, we consider the cell to be non-decodable 1 . The opportunistic algorithm works as follows.…”

Section: F Decoding Superimposed Encodingsmentioning

confidence: 99%

“…Key-value (k-v) storage has been used as a crucial component for many different network applications, such as social networks, online retailers, and cloud computing [1], [2]. Example implementations include Dynamo [3], Cassandra [4], Memcached [5], Redis [6], and BigTable [7].…”

Section: Introductionmentioning

confidence: 99%

kBF: A Bloom Filter for key-value storage with an application on approximate state machines

Xiong

Yao

Cao

et al. 2014

IEEE INFOCOM 2014 - IEEE Conference on Computer Communications

View full text Add to dashboard Cite

Abstract-Key-value (k-v) storage has been used as a crucial component for many network applications, such as social networks, online retailing, and cloud computing. Such storage usually provides support for operations on key-value pairs, and can be stored in memory to speed up responses to queries. So far, existing methods have been deterministic: they will faithfully return previously inserted key-value pairs. Providing such accuracy, however, comes at the cost of memory and CPU time. In contrast, in this paper, we present an approximate k-v storage that is more compact than existing methods. The tradeoff is that it may, theoretically, return a null value for a valid key with a low probability, or return a valid value for a key that was never inserted. Its design is based on the probabilistic data structure called the "Bloom Filter", which was originally developed to test element membership in sets. In this paper, we extend the bloom filter concept to support key-value operations, and demonstrate that it still retains the compact nature of the original bloom filter. We call the resulting design as the kBF (key-value bloom filter), and systematically analyze its performance advantages and design tradeoffs. Finally, we apply the kBF to a practical problem of implementing a state machine in network intrusion detection to demonstrate how the kBF can be used as a building block for more complicated software infrastructures.

show abstract

“…Note that many other challenges exist, including but not limited to the following: 1) making the storage stack much higher performance to take advantage of the low latency and high parallelism of raw flash devices [189] (similarly to what we discussed in Section 5.3 with respect to the Persistent Memory Manager), 2) providing transactional support in the flash translation layer for better system performance and flexibility [120], 3) taking advantage of application-and system-level information to manage flash memory in a way that improves performance, efficiency, lifetime and cost. These are great research directions to explore, but, for brevity, we will not discuss them in further detail.…”

Section: Flash Memory Scaling Challengesmentioning

confidence: 99%

Research Problems and Opportunities in Memory Systems

Mutlu

Subramanian

2014

JSFI

View full text Add to dashboard Cite

The memory system is a fundamental performance and energy bottleneck in almost all computing systems. Recent system design, application, and technology trends that require more capacity, bandwidth, efficiency, and predictability out of the memory system make it an even more important system bottleneck. At the same time, DRAM technology is experiencing difficult technology scaling challenges that make the maintenance and enhancement of its capacity, energyefficiency, and reliability significantly more costly with conventional techniques.In this article, after describing the demands and challenges faced by the memory system, we examine some promising research and design directions to overcome challenges posed by memory scaling. Specifically, we describe three major new research challenges and solution directions: 1) enabling new DRAM architectures, functions, interfaces, and better integration of the DRAM and the rest of the system (an approach we call system-DRAM co-design), 2) designing a memory system that employs emerging non-volatile memory technologies and takes advantage of multiple different technologies (i.e., hybrid memory systems), 3) providing predictable performance and QoS to applications sharing the memory system (i.e., QoS-aware memory systems). We also briefly describe our ongoing related work in combating scaling challenges of NAND flash memory.Keywords: memory systems, scaling, DRAM, flash, non-volatile memory, QoS, reliability. IntroductionMain memory is a critical component of all computing systems, employed in server, embedded, desktop, mobile and sensor environments. Memory capacity, energy, cost, performance, and management algorithms must scale as we scale the size of the computing system in order to maintain performance growth and enable new applications. Unfortunately, such scaling has become difficult because recent trends in systems, applications, and technology greatly exacerbate the memory system bottleneck. Memory System TrendsIn particular, on the systems/architecture front, energy and power consumption have become key design limiters as the memory system continues to be responsible for a significant fraction of overall system energy/power [112]. More and increasingly heterogeneous processing cores and agents/clients are sharing the memory system [11,36,39,60,78,79,178,181], leading to increasing demand for memory capacity and bandwidth along with a relatively new demand for predictable performance and quality of service (QoS) from the memory system [129,137,176].On the applications front, important applications are usually very data intensive and are becoming increasingly so [17], requiring both real-time and offline manipulation of great amounts of data. For example, next-generation genome sequencing technologies produce massive amounts of sequence data that overwhelms memory storage and bandwidth requirements of today's highend desktop and laptop systems [9,111,186,196,197] yet researchers have the goal of enabling low-cost personalized medicine, which requires even larger amoun...

show abstract

Using vector interfaces to deliver millions of IOPS from a networked key-value storage server

Cited by 24 publications

References 19 publications

kBF: Towards Approximate and Bloom Filter based Key-Value Storage for Cloud Computing Systems

kBF: Towards Approximate and Bloom Filter based Key-Value Storage for Cloud Computing Systems

kBF: A Bloom Filter for key-value storage with an application on approximate state machines

Research Problems and Opportunities in Memory Systems

Contact Info

Product

Resources

About