This Company is Helping Amazon Slash Server Requirements By 80%

The platform aims to improve the efficiency and cost-effectiveness of the tech giant's systems, resulting in a 1.6x increase in performance

In its hot chase for carbon-neutrality, Amazon Web Services (AWS) has teamed up with Aerospike, trying to close the gap with rivals Azure and Google Cloud. Aerospike, a multi-model, NoSQL real-time data platform, claims that its technology can assist Amazon in reducing its server requirements by 80% over three years. 

Additionally, the platform aims to improve the efficiency and cost-effectiveness of the tech giant’s systems, resulting in a 1.6x increase in performance and a reduction in the total cost of ownership. However, we often see companies make huge announcements in the direction of carbon-neutrality, but the projects never see the light of day. In fact, according to a recent survey conducted by the International Consumer Protection and Enforcement Network (ICPEN), a staggering 40% of companies are guilty of making false or misleading environmental claims.

Will Aerospike partnering with Amazon too be one of those projects? To gain more insights into this, Analytics India Magazine interacted with Aveekshith Bushan, vice president – Asia Pacific & Japan, at Aerospike

AIM Daily XO

Join our editors every weekday evening as they steer you through the most significant news of the day, introduce you to fresh perspectives, and provide unexpected moments of joy
Your newsletter subscriptions are subject to AIM Privacy Policy and Terms and Conditions.

Talking about how Aerospike’s combination with AWS Graviton2 impacts energy consumption and sustainability, Bushan says the company extracts significant efficiency from its hardware partners’ improvements in CPU, storage, and networking. According to a recent IEEE paper, Aerospike’s capacity to manage demanding workloads on clusters with lesser nodes benefits businesses looking to significantly lower their energy use.

“The reduced use of nodes,” said Bushan, “calls for lesser server requirements. Eliminating the physical infrastructure helps reduce carbon emissions and aids an organisation in achieving its sustainability goals.” According to Bushan, since fewer servers are needed because of the decreased node usage, the elimination of a part of the physical infrastructure also helps the company to achieve its sustainability goals. 

Download our Mobile App

Further, he presented a benchmark study on which the current standard for pricing performance is set – Aerospike on AWS Graviton – from Aerospike and AWS. 

“We basically ran a CPU-intensive workload with 300 clients connecting to Aerospike database version 6.2, which contained around 2 billion unique records, all stored in Aerospike’s in-memory storage option,” said Bushan, stating that the results were fantastic. 

While the benchmark clients were able to execute over 20 million read-only transactions per second, Aerospike processed 99% of them in under 1 millisecond. Bushan believes that this level of throughput and low latency is exactly what you need for high-demand applications like AdTech online bidding, fraud detection, payments processing, real-time personalization services and more.

There’s More 

Bushan also claimed that the resultant price-performance was 63% better than that of non-Graviton environments. Additionally, compared to other options, running Aerospike on AWS Graviton can even yield a significant reduction in carbon emissions. Bushan said, “AWS confirmed that by running an Aerospike workload on a Graviton cluster rather than a non-Graviton cluster for this benchmark scenario, it was possible to reduce carbon emissions by 49%.”

Bushan said that data production and consumption is growing at an unprecedented speed and many organisations weren’t prepared to handle it. “This poses a scaling challenge for the IT industry and a temporary solution is to update the infrastructure and server capacity to handle and process the massive volume of incoming data,” said Bushan. 

Aerospike uses a hybrid memory design where the in-memory index is not saved to the disk, but data is stored on persistent SSDs and read directly from the disk for fast access. With predictable performance, no disk I/O is needed to access the index. SSDs have the same read latency for both random and sequential access, allowing for optimisations to avoid costly device scans during index rebuilding.

In this benchmark, Aerospike kept all index and user data in memory (DRAM) — a configuration common for real-time fraud detection and online bidding applications in the AdTech industry. Other configuration options include storing all index and user data in PMem, on Flash (SSDs), or hybrid configurations with indexes in DRAM or PMem and user data on Flash.

“Such flexibility enables firms to use Aerospike in different ways to support different business needs without incurring excess hardware costs or compromising application requirements. It also helps firms standardise on a common database platform at the edge and at the core to simplify operations and reduce overall expenses,” concluded Bushan. 

Sign up for The Deep Learning Podcast

by Vijayalakshmi Anandan

The Deep Learning Curve is a technology-based podcast hosted by Vijayalakshmi Anandan - Video Presenter and Podcaster at Analytics India Magazine. This podcast is the narrator's journey of curiosity and discovery in the world of technology.

Lokesh Choudhary
Tech-savvy storyteller with a knack for uncovering AI's hidden gems and dodging its potential pitfalls. 'Navigating the world of tech', one story at a time. You can reach me at:

Our Upcoming Events

27-28th Apr, 2023 I Bangalore
Data Engineering Summit (DES) 2023

23 Jun, 2023 | Bangalore
MachineCon India 2023

21 Jul, 2023 | New York
MachineCon USA 2023

3 Ways to Join our Community

Telegram group

Discover special offers, top stories, upcoming events, and more.

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Subscribe to our Daily newsletter

Get our daily awesome stories & videos in your inbox