Microsoft Data Center Tour — Transcript

Explore Microsoft's global data centers, their design evolution, sustainability efforts, and how they deliver reliable cloud services worldwide.

Key Takeaways

  • Microsoft's data centers are designed for high efficiency, sustainability, and rapid scalability using modular components.
  • Renewable energy and innovative cooling technologies significantly reduce environmental impact and operational costs.
  • Global distribution and advanced network infrastructure ensure low latency, high availability, and geo-redundancy.
  • Security and operational excellence are top priorities, supported by continuous monitoring and multi-layered defenses.
  • Microsoft's cloud infrastructure supports a vast customer base with reliable, cost-effective, and sustainable services.

Summary

  • Microsoft's cloud is supported by a global network of data centers serving over a billion customers and 20 million businesses in 70+ countries.
  • The video tours multiple Microsoft data centers, highlighting their design, operation, and sustainability features.
  • Data center evolution from 1989 to present includes Generation 2, 3, and 4 facilities with increasing efficiency and modular designs.
  • Generation 2 data center in Quincy, WA, uses traditional raised floors, hot/cool aisles, UPS, emergency generators, and is powered by 100% renewable hydropower.
  • Generation 3 centers, such as Chicago and Dublin, introduced modular container units and economization cooling methods to reduce costs and energy use.
  • Water-side and air-side economization cooling significantly reduce electricity and water consumption compared to traditional data centers.
  • Generation 4 facilities use air-cooled, pre-manufactured IT packs for rapid deployment, reduced capital costs, and improved PUE (1.15-1.2).
  • Microsoft emphasizes sustainability, using renewable hydropower and recyclable materials in data center construction.
  • High availability and security are ensured via 24/7 operations centers, advanced monitoring, geographically distributed failover, and multi-layered security perimeters.
  • Microsoft's extensive fiber optic network and content delivery nodes reduce latency and improve service resiliency globally.

Full Transcript — Download SRT & Markdown

00:01
Speaker A
Where is the Microsoft cloud?
00:03
Speaker A
In reality, it exists in our vast worldwide network of data centers that provide the foundation for a wide array of online services.
00:12
Speaker A
When you use one of these services, for example, to do a search query on Bing, your request travels from your PC or mobile phone,
00:20
Speaker A
across the internet connected by an extensive global fiber optic network, through our content delivery network nodes, moving at the speed of light to a data center that executes the query and returns the results, all in a fraction of a second.
00:36
Speaker A
Microsoft's cloud is comprised of a globally distributed data center infrastructure supporting over 200 online services, more than a billion customers and 20 million businesses in over 70 countries use these services each year.
00:50
Speaker A
Today we're going to take you on a tour of a few of our data centers and show you how Microsoft's Global Foundation Services team,
01:02
Speaker A
the team that designs, builds, operates and secures our cloud infrastructure,
01:10
Speaker A
is delivering high availability and reliability, high efficiency, smart scalability, a secure, private, trusted cloud, and sustainability in all our facilities worldwide.
01:56
Speaker A
Microsoft has been operating data centers since 1989, and this is an area where Microsoft continues to make significant strategic investments.
02:05
Speaker A
From 1989 to 2004, our data centers were built to address early computer systems that required controlled temperature environments in which to operate.
02:14
Speaker A
From 2004 to 2007, we began to design our own large facilities to provide our online services with a range of solutions for more rapid deployment and more efficient operations.
02:27
Speaker A
In 2007, we opened our first Generation 2 data center in Quincy, Washington, which began to change the industry standards for large-scale internet data centers.
02:38
Speaker A
Today, the facility is approximately the size of 10 football fields and houses tens of thousands of high-performance processing and storage servers in high-density racks that are separated by hot and cool air aisles on a traditional raised floor.
03:10
Speaker A
Chillers and air handling equipment ensure precise control over environmental conditions.
03:39
Speaker A
Uninterruptible power supplies, or UPS systems, and vast banks of batteries ensure electricity remains continuous in the event of a short-term power disruption.
03:51
Speaker A
Emergency generators provide backup power for extended outages and for planned maintenance and can operate the data center with onsite fuel reserves in the event of a natural disaster.
04:02
Speaker A
A high-speed robust fiber optic network connects this data center with other major hubs and internet users.
04:20
Speaker A
Edge compute nodes host workloads closer to the end users to reduce latency, provide geo-redundancy, and increase overall service resiliency.
04:59
Speaker A
A team of engineers work around the clock to help ensure services are persistently available to customers.
05:07
Speaker A
This facility is powered 100% by renewable hydropower.
05:15
Speaker A
To deliver cloud computing services for global businesses and consumers within a very cost-effective and sustainable operation, we evolved our data center design strategy.
05:33
Speaker A
Traditional data centers take from 18 to 24 months to build, with our modular approach, we work to reduce costs and deployment time while increasing overall efficiency and sustainability.
05:49
Speaker A
In 2009, we opened our Chicago data center that spans approximately 10 football fields in size.
06:01
Speaker A
It is a Generation 3 facility and was our first deployment of modular computing units to dramatically reduce infrastructure costs
06:12
Speaker A
and the time to deploy large volumes of compute power.
06:22
Speaker A
We used standard shipping containers to house up to 2,400 servers that were trucked onsite and plugged into power, water,
06:40
Speaker A
and networking infrastructure already in place.
06:44
Speaker A
This enabled us to scale out capacity to meet customer demands for services within hours versus weeks and significantly reduced packaging materials,
06:55
Speaker A
transportation waste, and carbon emissions.
07:00
Speaker A
Another green IT approach is the use of water-side economization, which enables us to cool the facility without requiring the high levels of electricity typically needed to power large chillers.
07:09
Speaker A
Power Usage Effectiveness or PUE is the measure of total data center power consumption divided by IT or critical power,
07:24
Speaker A
and highlights how much power is consumed by supporting infrastructure versus the power to operate servers.
07:36
Speaker A
A typical industry data center has a PUE of approximately 2.0, where the amount of energy consumed for the facilities and supporting infrastructure
07:49
Speaker A
equals the energy to power the servers.
07:55
Speaker A
This award-winning data center has a PUE from 1.15 to 1.22 for the containers in the facility.
08:04
Speaker A
Also in 2009, we further developed the containment approach to deliver additional efficiencies in our Dublin Generation 3 facility.
08:20
Speaker A
This facility spans approximately seven football fields in size and deploys large-scale compute capacity contained in pods of servers
08:30
Speaker A
cooled by air-side economization.
08:34
Speaker A
Air handling units on the roof draw in cold outside air to provide the cooling for the servers and facility.
08:43
Speaker A
Evaporative coolers, vaporizing water into the air to absorb heat, and an air bypassing feature improves operational efficiency and maintains constant room temperature regardless of outdoor conditions.
08:58
Speaker A
The facility uses less than 1% of the annual water consumption of a traditional data center.
09:09
Speaker A
Our Dublin facility maintains a PUE of 1.25 and improves energy efficiency by approximately 50% as compared to our traditional data centers of similar capacity.
09:22
Speaker A
In 2010, we began to build our first Generation 4 data center design, building on the Dublin experience of using air-side economization into a truly modular design.
09:34
Speaker A
This technologically advanced facility challenged every industry design standard to significantly reduce water and power use and it uniquely demonstrates how we have evolved our approach.
09:50
Speaker A
Our Generation 4 facility utilizes air-cooled IT pre-manufactured components or IT packs.
10:00
Speaker A
These state-of-the-art pre-assembled plug-and-play modular components can be manufactured in many locations through a dynamic supply chain of globally distributed vendors
10:16
Speaker A
and use recyclable materials such as steel and aluminum, which are then shipped onsite and assembled.
10:25
Speaker A
IT packs allow rapid commissioning of additional compute capacity as the business needs it, reducing the typical two-year construction timeframe in half
10:38
Speaker A
and reducing capital costs by 30 to 50% over the lifetime of the facility.
10:46
Speaker A
The IT packs also provide self-contained UPS, along with air handling and evaporative cooling systems to control ambient temperatures with four cooling modes to ensure the inlet temperature is maintained between 50 and 90 degrees Fahrenheit
11:00
Speaker A
and 20 to 80% relative humidity.
11:05
Speaker A
Our simple, recyclable metal building provides shelter from weather and directs proper airflow through all systems.
11:14
Speaker A
This approach enables us to deliver PUEs of 1.15 to 1.2 across the facility and energy is provided from 100% renewable hydropower.
11:25
Speaker A
Operational excellence and services reliability comprise our highest priorities.
11:36
Speaker A
Microsoft ensures high availability through advanced monitoring and incident response, service support, and backup failover capability
11:48
Speaker A
managed through our geographically distributed Microsoft Operations Centers, operating 24x7x365.
11:56
Speaker A
We also deploy one of the largest implementations of Microsoft System Center in the world, and we use our own data center repository called Scry
12:09
Speaker A
to record power consumption and allow precise cost allocation to our internal business groups.
12:18
Speaker A
In addition, our network is one of the largest in the world, with a fiber optic and content distribution network backbone
12:29
Speaker A
connecting our data centers and edge nodes to ensure high performance and reliability.
12:35
Speaker A
A large geographically distributed footprint of data centers enables us to be close to customers to reduce network latency
12:45
Speaker A
and allow for geo-redundant backup and failover.
12:50
Speaker A
Security at our data centers employs outer and inner perimeters with increasing security at each level,
13:00
Speaker A
utilizing a combination of technology and traditional physical measures.
13:08
Speaker A
Technical elements include two-factor access control, badge readers, extensive camera monitoring,
13:16
Speaker A
and integrated alarm systems.
13:20
Speaker A
Traditional measures can include perimeter fencing, security officers, and locked server racks.
13:28
Speaker A
Data secured in Microsoft's data centers is classified based on its value and protected accordingly.
13:36
Speaker A
For example, hard drives with highly sensitive information are routinely destroyed when they are decommissioned.
13:45
Speaker A
We also limit access to our facilities and servers to only those individuals who have valid business needs.
13:53
Speaker A
In addition to maintaining best-in-class operations processes, we routinely bring in third parties to verify our capabilities,
14:05
Speaker A
resulting in ISO and PCI certifications, SAS 70 Type II attestations,
14:13
Speaker A
and FISMA certification and accreditation.
14:17
Speaker A
Microsoft's award-winning cloud infrastructure powers the cloud services our customers and partners use every day.
14:26
Speaker A
So the next time you search on Bing, expand your customer services with Windows Azure, communicate with employees via Office 365,
14:38
Speaker A
email your family on Hotmail, or challenge your friends across the world on Xbox Live,
14:48
Speaker A
you too are tapping into the power of the Microsoft Cloud.
Topics:Microsoft data centercloud infrastructurerenewable energydata center coolingmodular data centerssustainabilityhigh availabilitycloud computingfiber optic networkdata center security

Frequently Asked Questions

What powers Microsoft's data centers sustainably?

Microsoft's data centers are powered 100% by renewable hydropower, reducing their carbon footprint and supporting sustainable operations.

How does Microsoft ensure high availability in its data centers?

Microsoft uses 24/7 operations centers, advanced monitoring, geographically distributed failover, and backup power systems to maintain high availability.

What innovations have Microsoft introduced in data center design?

Microsoft evolved from traditional designs to modular, pre-manufactured IT packs with advanced cooling methods, reducing deployment time, costs, and energy use.

Get More with the Söz AI App

Transcribe recordings, audio files, and YouTube videos — with AI summaries, speaker detection, and unlimited transcriptions.

Or transcribe another YouTube video here →