Twitter Boosts Performance and Cost Efficiency

Twitter increases Hadoop performance and cost efficiency with caching, fast SSDs and more compute.

At a Glance:

  • Twitter uses Hadoop* for storing data and performing advanced analytics to generate important business insights.

  • Twitter expects that caching temporary data with Intel® SSDs based on Intel® 3D NAND Technology and increasing core counts with 2nd Gen Intel® Xeon® Scalable processors will result in approximately 30 percent lower TCO and over 50 percent faster runtimes, compared to their legacy production cluster configuration.1

author-image

Par

Executive Overview

Storage I/O can be a significant performance bottleneck for Hadoop* clusters, especially in hyperscale deployments like those at Twitter, where a single cluster can have up to 10,000 nodes and nearly 100 PB of logical storage. The typical Hadoop cluster at Twitter contains over 100,000 hard disk drives (HDDs)—but this configuration was reaching an I/O performance limit because while HDD capacity has increased over time, HDD performance has not significantly changed.2 Therefore, simply adding more, bigger HDDs wasn’t going to solve Twitter’s scaling challenges—in fact, it would make things worse as the I/O per GB decreases. Adding more spindles per node was not feasible due to space and power limitations.

Working in collaboration with an Intel engineering team, Twitter engineers conducted a series of experiments that revealed that storing temporary files managed by YARN* (Yet Another Resource Negotiator*) on a fast SSD enabled significant performance improvements on existing hardware (up to a 50 percent reduction in runtime).3 The team also discovered that removing a storage I/O bottleneck enabled them to use larger hard drives while simultaneously increasing processor utilization, which in turn resulted in the ability to use higher-core-count processors. This positively affected storage performance, and contributed to higher data center density by reducing the number of required HDDs.

Higher density leads to total cost of ownership (TCO) savings through energy efficiency, fewer racks, and a smaller data center footprint. Overall, Twitter expects that caching temporary data and increasing core counts will result in approximately 30 percent lower TCO and over 50 percent faster runtimes, compared to their legacy production cluster configuration.1

Read the white paper - Boosting Hadoop* Performance and Cost Efficiency with Caching, Fast SSDs, and More Compute

Download the PDF ›

Découvrez les produits et solutions associés

Infos sur le produit et ses performances

1

Base de référence : processeur Intel® Xeon® E3-1230 v6 à un socket (4 cœurs) ; 32 à 64 Go de RAM ; 1 disque dur de 1 To ou 2 To ; disque de démarrage Intel S4500 de 240 Go ; carte Ethernet de 1 GbE à 10 GbE ; pas de mise en cache. Test : processeur Intel® Xeon® Gold 6262 à un socket (24 cœurs) ; 192 Go de RAM ; disque de démarrage Intel S4500 de 240 Go ; 8 disques durs de 6 To ; 1 unité de stockage SSD Intel® DC P4610 de 6,4 To ; carte Ethernet 25 GbE ; mise en cache à l'aide du logiciel Intel® Cache Acceleration Software (Intel® CAS). Système d'exploitation : Twitter CentOS* 6 Derivative, version du noyau 2.6.74-t1.el6.x86_64 (basé sur le noyau 4.14.12 en amont), version du BIOS : D3WWM11, version du microcode : 0xb000021.

2

Backblaze, septembre 2018, « Hard Disk Drive (HDD) vs Solid State Drive (SSD): What’s the Diff? » (Disque dur vs lecteur SSD : quelle est la différence ?) https://www.backblaze.com/blog/hdd-versus-ssd-whats-the-diff/.

3

Référence : processeur Intel® Xeon® E5-2630 v4 à deux sockets à 2,2 GHz (10 cœurs/20 threads par socket) ; 128 Go de RAM ; 12x disques durs SATA de 6 To à 7200 tr/min ; 1 lecteur de démarrage SSD SATA ; carte Ethernet 25 GbE ; 102 nœuds répartis sur 6 racks. Charge de travail : Gridmix* et Terasort*. Score Gridmix : 3309 secondes ; score Terasort : 5504 secondes ; test : processeur Intel® Xeon® E5-2630 v4 à deux sockets à 2,2 GHz (10 cœurs/20 threads par socket) ; 128 Go de RAM ; 12 disques durs SATA de 6 To à 7200 tr/min ; 1 lecteur de démarrage SSD SATA ; 1 lecteur SSD Intel® Optane™ DC P4800X NVMe* ; carte Ethernet 25 GbE ; 102 nœuds répartis sur 6 racks. Charge de travail : Gridmix et Terasort. Score Gridmix : 2396 secondes ; score Terasort : 2640 secondes ; SE : Twitter CentOS* 6 Derivative, noyau.