By Jenna Boller, Vice President of Marketing
As we witness the exponential growth of data + AI across nearly every industry, data centers — both on premises and in the cloud — have been tasked with processing increasingly vast amounts of data. Yet, these data centers present environmental challenges. Over the past several years, I’ve seen a growing number of headlines expressing concern about the environmental impact of data centers. This trend has only accelerated with the boom of GenAI technologies over the past 18 months.
While Big Tech companies like Microsoft, Google, and Amazon are building nuclear and geothermal power plants to power their energy-hungry data centers, it’s unlikely that a typical enterprise running workloads on premises can pursue such options and gain efficiencies of scale at the same level as the largest cloud services providers.
Ocient, as both a company and a technology platform, was founded with a mission to enable previously infeasible price-performance for always-on, compute-intensive data analysis. Through our diligent focus on maximizing hardware and software efficiency, we’ve discovered that Ocient’s approach to optimizing for performance and cost has translated into reducing our high-performance software solutions’ system footprints and energy consumption.
As I discuss how Ocient can help enterprises and government agencies tackle their energy reduction goals, I’ll explore:
- The hype behind the growing environmental impact of data center growth on the world’s energy resources.
- The need to find ways to reduce — not just offset — the energy data centers consume.
- Ways in which Ocient is leading the charge to deliver more efficient software solutions that can support goals related to carbon reduction and sustainable IT.
Headlines pulled from the Washington Post, Newsweek, and SiliconAngle.
While data center energy consumption is not a new topic in the media, a growing number of headlines point to concerns about how much the current Gen AI boom will affect the planet and our most critical resources. From water to energy, data centers are consuming more resources every day. In fact, organizations like Climatiq report that cloud computing emissions now exceed those from commercial aviation.
With data center energy consumption expected to consume 10% of the world’s energy supply by 2030, companies must find ways to reduce their impact.
AI consumes orders of magnitude more energy
On top of growing data center electricity usage, we are now seeing added spikes from LLMs and Gen AI. According to the International Energy Agency, “Search tools like Google could see a tenfold increase of their electricity demand in the case of fully implementing AI in it. When comparing the average electricity demand of a typical Google search (0.3 Wh of electricity) to OpenAI’s ChatGPT (2.9 Wh per request) and considering 9 billion searches per day, this would require almost 10 TWh of additional electricity in a year.”
Cryptocurrencies also contribute substantially to global data center energy usage, but “traditional data centers” will continue to pull a majority share of energy from the grid.
Global Data Center Electricity Use Will Outpace Entire Countries in ~20 Years
In case the above charts weren’t enough, as of 2021, global data center electricity usage was equal to all the electricity consumed by the state of California (pop: 39 million). By 2028, data centers are expected to consume as much energy as the entire country of Brazil (pop: 217 million). That’s massive growth, which will outpace the supply of clean energy sources and potentially swamp existing “traditional” sources of electricity.
Carbon offsets and clean energy sources won’t solve this problem alone
So, it’s not just hype. The energy consumed by data centers, especially when it comes to resource-intensive machine learning and AI use cases, is massive and growing. This problem is exacerbated by the fact that energy resources are limited. Data centers draw from the same grids as vehicles, households, and businesses, pushing utility companies to the breaking point even as they prepare for a global energy transition.
As more organizations look to utilize clean energy, these alternative sources are struggling to meet demand. Carbon offsets offer an approach to balancing the carbon emissions of intensive data center energy usage, but they don’t always prove to be effective and can be harder to track.
Software must be simply more efficient
For CIOs, CTOs, data center managers, and anyone concerned with their organizations’ carbon reduction goals, it’s imperative to look for ways to offset greenhouse gas emissions while leveraging clean energy sources. But this is not enough – organizations must also find ways to reduce their overall energy consumption.
A few years ago, this may have been seen as a “nice to have.” Today, it’s a growing imperative in the United States and many countries abroad.
In the United States, the U.S. Securities and Exchange Commission (SEC) now requires that “climate risk disclosures be included in a company’s SEC filings, such as annual reports and registration statements.” This includes following the Greenhouse Gas (GHG) Protocol to disclose Scope 1 and Scope 2 emissions.
In Europe last year, the Energy Efficiency Directive adopted by the European Parliament and Council established a legally binding target to reduce the EU’s energy consumption by 11.7% by 2030. The directive includes “an obligation for the monitoring of the energy performance of data centres.”
Reducing energy consumed by data centers can help decrease an organization’s Scope 2 greenhouse gas emissions. This works across the entire supply chain, so companies and public sector agencies that reduce energy consumption for “data as a service” products and offerings may also help their stakeholders and consumers reduce Scope 3 emissions as part of the overall tech, intelligence, and IT supply chain.
This journey begins and ends with making it a priority to require any new software solutions under evaluation to model their energy usage for the near term, medium term, and long term. IT departments and business users should request transparency from their IT vendors so that they have a handle on their supply chain and the greenhouse gas emissions resulting from their technology stack.
Ocient provides powerful data processing, analytics, and machine learning capabilities while reducing system footprint and energy consumption by 50% to 90%
This is where Ocient comes in. As an innovative data analytics software solutions company, Ocient offers a strategic software solution for CIOs, CTOs, and IT business leaders aiming to reduce energy consumption within their data center and cloud deployments, particularly for always-on, compute-intensive workloads.
Ocient specializes in managing large-scale datasets (up to trillions of records and petabytes of active data under analysis) and excels in geospatial data analytics, machine learning, SQL data warehousing, data integration, and real-time analytics. Traditional management of such large-scale datasets requires vast computational power, which results in significant carbon emissions. However, Ocient’s software solutions provide a far more efficient approach — up to 90% more efficient than competitors — reducing overall system footprint, cost, and energy consumption.
The underlying tech driving Ocient’s energy reduction outcomes
Ocient’s size, cost, and energy efficiency benefits stem from our breakthrough architectural approach to facilitating always-on, compute-intensive data analysis on large-scale datasets.
While optimizing for efficient price-performance for resource-intensive data analytics workloads, Ocient developed several unique software architectural innovations, including:
Ocient’s Compute Adjacent Storage Architecture™ (CASA)
With a revolutionary architecture that places compute adjacent to storage, CASA removes the remote object storage tier leveraged by cloud data warehouses, eliminating network bottlenecks and enabling rapid data access. CASA pushes down as much data processing as possible into the I/O layer of Ocient’s software solution stack, delivering 420x more data bandwidth in 1/10th the number of drives (or cloud instances) as a remote object storage data warehouse architecture.
Zero Copy Reliability™
Ocient leverages forward error correction and parity coding to ensure reliability and accuracy without 2x-3x data replication. While other databases and data warehouses copy data for reliability, Ocient uses Zero Copy Reliability to minimize the system footprint and maximize hardware resources and cloud instances.
TimeKey® clustering and a comprehensive set of indexes
A comprehensive set of indexes, including TimeKey clustering, geospatial, hash, n-gram, and inverted indexes, boosts query performance by up to 300x with very little capacity overhead.
Ocient Megalane™
Ocient Megalane facilitates compute-intensive workloads by enabling an extreme number of parallel tasks, using standard industry hardware and cloud instances to their full potential.
Native support for ETL/ELT, AI/ML, real-time analytics, and geospatial data analytics
Ocient enables users to do more than SQL data warehousing within a single solution, reducing the need for complexity and multiple-point solutions that can drive up energy usage for large datasets.
Read more about the Ocient Hyperscale Data Warehouse™ in our tech paper here.
Ocient provides deployment flexibility to run on 100% renewable energy
In addition to providing a more efficient software solution for hyperscale data analytics on premises and in Google Cloud or AWS, Ocient gives users the option to deploy into OcientCloud® — fully managed and operated by Ocient — which runs on 100% renewable energy. This enables users to further reduce concerns around GHG emissions resulting from workloads running on Ocient.
Run more efficient data analytics solutions that reduce energy, space, and cost by up to 90%
To bring Ocient’s disruptive software solutions and core architectural innovations into perspective, I’d like to further demonstrate the overall sizing, footprint, and energy consumption benefits of running Ocient in your data center when compared to other incumbents on the floor.
Ocient deployment compared to a NoSQL on-premises database
As part of a compliance-focused use case, one of our telecommunications customers needed to migrate their 5-petabyte workload from a legacy on-premises database solution to a modern data warehouse or database. In comparing Ocient to a NoSQL database, Ocient demonstrated the ability to run the same workload on 12 nodes versus a 170-node NoSQL system.
Thanks to Ocient’s columnar database architecture, compression, Zero Copy Reliability, and Compute Adjacent Storage Architecture, the Ocient system requires only 12Kw of energy to run versus 115Kw. This equates to a 902,280 KWh energy savings per year for an always-on workload running at peak power.
Ocient deployment compared to a legacy Hadoop deployment
The same customer was looking to replace Hadoop on the floor of their data center. Running the same workload, with better performance, required 12 nodes of Ocient versus 104 nodes from the Hadoop system. When deployed in the data center, this translated to 7 racks of Hadoop versus ½ a rack of Ocient.
Average power consumption for Ocient remained at 12Kw versus 112Kw for the Hadoop system.
In short, Ocient is not just a software solutions company but a key player in the drive for more sustainable data management at scale. Ocient provides a sustainable software solution that CTOs, CIOs, IT business leaders, and organizations at large should consider as they look for ways to reduce the energy consumption and environmental impact of their data management solutions. As part of this journey, we’re proud to have recently announced that we are backed by Buoyant Ventures, a venture fund that is just as passionate about decarbonizing data centers as we are.
For businesses with large data needs spanning telecommunications, public sector, advertising technologies (AdTech), climate risk analytics, vehicle telematics, energy utilities, financial institutions, airlines, automotive companies, and more, Ocient is the strategic choice to support the decarbonization of your data center emissions and overall carbon reduction goals. By significantly reducing the price to run complex computations at scale, Ocient enables companies to meet their data needs and make significant progress on their climate action goals.
Get a free energy and TCO savings analysis
Curious to see the potential cost, sizing, and energy savings Ocient can deliver for your workloads? Get in touch for a FREE energy and TCO savings analysis, and let’s lower the energy consumed by high-powered, high-intensity data analytics together.