Making Search Engines Faster by Lowering the Cost of Querying Business Rules Through FPGAs

Maschi, Fabio; Owaida, Muhsen; Alonso, Gustavo; Casalino, Matteo Maria; Hock-Koon, Anthony

doi:10.1145/3318464.3386133

Cited by 8 publications

(19 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 1 shows a simplified, but syntactically representative example of how the MCT rules look like (the actual rules have thirty-four criteria). We refer to [15] for further information on how the MCT module processes the rules.…”

Section: Mct: Filtering Impossible Connectionsmentioning

confidence: 99%

“…Figure 2 depicts erbium [15], the NFA-based Business Rule Engine hardware accelerator. Its different elements can be decomposed into offline and online modules.…”

Section: Erbium Enginementioning

confidence: 99%

“…While some works present the hardware efficiency vis-à-vis the cloud infrastructure [2,4,7], and the vast FPGA literature emphasises kernel acceleration under a stand-alone context, very few analyse the deployment efficiency of applying FPGAs in real computing systems. In this paper, we present one such study, based on recent research results that demonstrated significant potential gains when implementing part of a search engine on an FPGA [15,17]. The initial prototype was turned into a Proof-of-Concept deployment tested on the real system and computing infrastructure, as well as evaluated against real data and under the constraints imposed by the existing search engine.…”

Section: Introductionmentioning

confidence: 99%

“…An initial implementation of MCT on top of an FPGA proved to be a significant improvement over the existing system along several dimensions [15]. By using a Non-deterministic Finite State Automaton (NFA) and exploiting the inherent parallelism and pipeline possibilities available on the FPGA, the performance gains over the existing system were impressive.…”

Section: Introductionmentioning

confidence: 99%

“…The downtime for updating the rules is only 500 𝜇s, four orders of magnitude faster than in the current system, which can be used to improve the overall availability of the search engine. Moreover, the performance gains indicated that the improvements could be exploited in several ways in addition to increasing throughput and reducing latency: the quality of the search could be improved by considering more options as part of the search; the computing capacity needed for the search engine could be reduced; and the architecture became more flexible, offering several ways to integrate the MCT and other components (see [15,17] for more details and an extensive discussion of the deployment possibilities of FPGAs within the flight search engine).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

From Research to Proof-of-Concept: Analysis of a Deployment of FPGAs on a Commercial Search Engine

Maschi,

Alonso,

Hock-Koon

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

FPGAs are quickly becoming available in data centres and in the cloud as a one more heterogeneous processing element complementing CPUs and GPUs. There are many reports in the research literature showing the potential for FPGAs to accelerate a wide variety of algorithms, which combined with their growing availability, would seem to also indicate a widespread use in many applications. Unfortunately, there is not much published research exploring what it takes to integrate an FPGA into an existing application in a cost-effective way and keeping the algorithmic performance advantages. Building on recent results exploring how to employ FPGAs to improve the search engines used in the travel industry, this paper analyses the end-to-end performance of the search engine when using FPGAs, as well as the necessary changes to the software and the cost of such deployments. The results provide important insights on current FPGA deployments and what needs to be done to make FPGAs more widely used. For instance, the large potential performance gains provided by an FPGA are greatly diminished in practice if the application cannot submit request in the most optimal way for the FPGA, something that is not always possible and might require significant changes to the application. Similarly, some existing cloud deployments turn out to use a very imbalanced architecture: a powerful FPGA connected to a not so powerful CPU. The result is that the CPU cannot generate enough load for the FPGA, which potentially eliminates all performance gains and might even result in a more expensive system. In this paper, we report on an extensive study and development effort to incorporate FPGAs into a search engine and analyse the issues encountered and their practical impact. We expect that these results will inform the development and deployment of FPGAs in the future by providing important insights on the end-to-end integration of FPGAs within existing systems.

show abstract

Section: Mct: Filtering Impossible Connectionsmentioning

confidence: 99%

“…Figure 2 depicts erbium [15], the NFA-based Business Rule Engine hardware accelerator. Its different elements can be decomposed into offline and online modules.…”

Section: Erbium Enginementioning

confidence: 99%