From High-Performance to Hyperscale
Data Center Case Study
Since the 1960s, data centers have experienced a rapid evolution that continues to shift every day. What began as an 1,800 square-foot, 30-ton machine is now a variety of compact computer servers in dedicated buildings known as data centers. Today, data is delivered through a variety of outlets, from small high-performance computers to mega-sized hyperscale data centers.
“Over the next 10 years, we’ll reach a point where nearly everything has become digitized.”
CEO of Microsoft
DATA CENTER TYPES
What does Hyperscale Computing mean?
Hyperscale computing refers to the facilities and provisioning required in distributed computing environments to efficiently scale from a few servers to thousands of servers. Hyperscale computing is usually used in environments such as big data and cloud computing.
The structural design of hyperscale computing is often different from conventional computing. In the hyperscale design, high-grade computing constructs, such as those usually found in blade systems, are typically abandoned. Hyperscale favors stripped-down product design that is extremely cost effective. This minimal level of investment in hardware.
Cisco estimates that by 2021, traffic within hyperscale data centers will quadruple, and hyperscale data centers will account for 55% of all data center traffic by 2021. This growth will only continue as the hyperscale data center market is expected to reach $80.65 billion by 2022, according to Markets and Markets.
source: equinix blog
Overcoming Data Center Challenges
The construction of data centers requires an elevated level of safety compared to other projects. In addition to the typical hazards of construction, teams are often working around live systems as they construct and connect electrical building components. With such sensitive systems and potential electrical and chemical hazards, contractors and designers are trained to understand the “why” behind every design and activity in construction. Often times, designs are changed to better mitigate any safety risk.
SAFETY RISKS INCLUDE:
- Working on an operating campus
- Working in a critical environment
- Managing accelerated activities and overtime
- Ground disturbance
- Live electrical environments
- Hazardous materials
- Heavy equipment
- Trade stacking
- Inner trade electrical stacking
- Flammable materials
- Working from heights
- Lifting operations
- Confined spaces
Mission Critical Focus
Though our safety program is instrumental on every project, we take an extra step in implementing project specific practices on data centers. These projects have sensitive systems and potential electrical and chemical hazards every team member is trained to understand and work around. We emphasize:
Additional Safety Obstacles
Because many data centers fall under strict NDA’s, preventable data center incidents are often not shared with the community.
An evergrowing concern for data centers is the cost of energy consumption. U.S. data centers use more than 90 billion kilowatt-hours of electricity a year. Global data centers used roughly 416 terawatts (4.16 x 1014 watts) (or about 3% of the total electricity) in 2016. And this consumption is projected to double every four years. (source: Forbes.com)
The rising energy costs and rapidly growing data usage have placed the onus of energy efficiency on data center clients. Our Clayco team utilizes green building practices to deliver energy efficient buildings that provide our clients with a more optimal power usage effectiveness (PUE) ratio.
POWER DEMANDS OF DATA CENTER TYPES
Moderate Power Demand
High powered computer typical in science applications and as of recently, trading.
Specs:. Large Power Requirements (typ) Water Cooled | Applications to support Artificial Intelligence (AI) | Batch Processing, Algorithm Calculations, Mapping, etc. | No redundancy requirements
Markets include: Government | Science | University | Trading
Moderate Power Demand
Colocation data centers located near a Telecom Central Office for lower latency. Often used for IoT applications.
Specs:. Near Telecom Central Offices | Low Latency applications i.e. GPS, camera, search engines, gaming, etc | Supports IoT (Internet of Things), connectable to core data centers | 750 kVA-2MW standard pod
Markets include: GPS | Camera | Search Engines | Gaming
Colocation Data Center
Substantial Power Demand
Builds and leases on an individual rack or room of racks basis. Typical suites are 10,000 square feet or 2MW.
Specs: Sales quotas typ. 175kVA per quarter | Racks $600-$1200/Month | Managed Services | Catcher Block System (13 to make 12 blocks) | Caged Environments | Vertical Market Sales (legal, healthcare, financial, etc.) | Targeting IoT, DRaas, and migrating towards Edge
Markets include: Colocation providers (i.e. Equinix)
Wholesale Data Center
Substantial Power Demand
Provides Data Center Suites (2MW+) to Enterprise and Hyperscale clients. Leases racks and offers managed services.
Specs:85%+ Enterprise Market | Large Campuses | Current Focus: Hyperscale Data Centers | Typ. 150-200 watts/SF | Largest Markets: Ashburn, Virginia, Dallas, and Chicago | Basis of ESD Colo4Hyperscale Design
Markets include: Wholesale Colocation providers (i.e. Cyrus One, Iron Mountain) | (Photo Source: ESD)
Enterprise Data Center
Extensive Power Demand
Data processing that supports operation i.e. retail, manufacturing, healthcare, and finance.
Specs: 85% Outsourced to Collocation/Wholesale | Cloud Applications = Decreased Loads | Sold as Sales/Leaseback | Previously Uptime
Markets include: Retail | Manufacturing | Healthcare | Finance
Hyperscale Data Center
Extensive Power Demand
Software Developer offering product within the cloud. Primary business is software development or processing based workload.
Specs:Critical Network | Economy Servers | 85-95 Degrees Inlet Temp | Evaporative Cooling | 99.999 Reliability (typ. Tier 2+) | 75% Operating Distributed Redundant Design | Dual MDF/4 IDF | 25-32MW Tranche’s
Markets include: Google | Amazon | Facebook | Apple | Microsoft
Building Community Craft
Most hyperscale data centers are located in rural areas, usually hours away from the larger workforces that would be available to a jobsite in an urban area. This presents the issue of labor on data center project construction. Finding good talent in remote locations takes a unique approach. Clayco’s commitment to the communities we work in allows us to build relationships with local contractors and subcontractors. Our company builds a project-based program to mentor new labor, taking time to teach the aspiring workforce the intricate details required to build safe and optimally functioning facilities. Because data centers are complex, technical facilities, these projects are opportunities for the local labor force to increase their expertise.
Craft retention is also of utmost importance, sometimes a challenge in rural areas on complex projects. Our safety program, procurement process and jobsite culture are all focused on attracting and retaining the strongest trade contractors onsite. Workers must have adequate access to clean and supplied wash room facilities, break areas, a clean and safe work place and an opportunity to communicate with our onsite staff, voicing any concerns or additional needs. Workers must feel they are an important part of the team.
These large data center projects should positively impact every community they touch.