facebook-pixel
  • Data Science
  • Hossein Eslambolchi
  • MAY 04, 2018

Datacenter Evolution and Implications

Ready to learn Data Science? Browse courses like Data Science Training and Certification developed by industry thought leaders and Experfy in Harvard Innovation Lab.

1.0 INTRODUCTION

One of the most significant technological breakthroughs in recent memory is that of the cloud. It is hard to imagine the world without “the cloud” these days as it has permeated just about every market one can think of. Both public and private clouds have quickly become a business necessity with the explosion in the availability of data in recent years.

Cloud services now provide the capability to both store and compute vast amounts of data. The digitization of media consumption has also greatly catalyzed the growth of cloud services. Over-the-top (OTT) services such as Spotify and Netflix have benefited from cloud technology which allows them to not only serve media to their customers but also to compute and analyze vast amounts of harvested customer data.

You hear a lot about so-called “Big Data” but I think this term is a bit too conservative – indeed I prefer the term “Super Massive Data”. Businesses all over the world have begun to rely on cloud technology to address this massive amount of data and to meet their customer’s needs and wants. And it is not the only business that thrives with the advent of this breakthrough technology – academia, governments, and consumers all reap the benefits too.

Indeed, cloud computing is widely leveraged across a variety of problem domains ranging from movie recommendation systems to unraveling the mysteries of the universe. Consider the Search for Extraterrestrial Intelligence project (SETI) which seeks to find extraterrestrial life beyond our planet. The amount of data to be analyzed is truly astronomical! The SETI@home project out of Berkeley stands out as a particularly unique approach to harvesting distributed computing power. This concept of a distributed network of computers is foundational to the modern notion of the cloud.

2.0 THE FEDERATED CLOUD

Besides the shared reliance on cloud technology, no two companies share the same exact set of business requirements. As such, a one-stop-shop cloud service is often not sufficient for all businesses. Enter the idea of the Federated Cloud – a hybrid cloud composed of both private and public clouds from one or more providers. A Federated Cloud allows a business to tailor their cloud platform to their specific needs.

3.0 THE INFRASTRUCTURE

Building a robust cloud architecture is complicated and challenging work. For example, just consider the challenge of connecting the continents with undersea cables – a crucial component of a properly distributed cloud architecture. As the internet continues to spread to the remote regions of the world the internet will soon follow suit. It’s estimated that only 47 percent of the world population use the internet.

The cloud of the 21st century must be built to support the remaining 43 percent. But even with the current 47 percent coverage, consider too that in the US only about 3 percent of the fiber infrastructure is currently utilized – an astounding 97% remains to be lit! As fiber slowly becomes the norm, cloud services will need to greatly expand capacity. From experience, I have observed that data center expenses are directly proportional to the number of users being supported.

For example, data center operational expenses include infrastructure, servers, power, HVAC – not to mention the dollar/GB bandwidth costs of delivering content to the user! There’s also the required IT and personnel required to maintain the infrastructure. By my estimates, for every 1 million users, data center expenses and network complexity increases by a factor of 10. With current cloud infrastructure supporting approximately 3 billion people today, the thought of supporting all 7.5 billion people on earth is truly massive.  

In the early days of the internet in 1969, no one could ever have imagined the enormous impact that the internet has today. Back then, the internet was used to send e-mail messages between two computers and was designed with US defense squarely in mind. It was not until the mid-1990’s during President Clinton’s term when the internet started taking off on a massive scale through web-based services. Since then, the amount of data has exploded with richer and richer content. The Tokyo 2020 Summer Olympics are expected to be broadcast in 8K resolution. Even if we continued to only support the 3 billion users of today, the rate at which content is improving and expanding is mind blowing. VR and AR technology continue to improve and the infrastructural support required will only increase.

Security is another major challenge for modern clouds. Just as in those old western films depicting a bank robbery or a train heist, a large part of cyber crime today involves social engineering and understanding business operations – when they open and close, the number of security guards at any given time, etc. Black hat hackers utilize any means necessary to infiltrate and exploit cloud weaknesses.

Loss of intellectual property is a chief concern for CEOs at corporations around the world. It is also important to understand that cloud services are not immune to DDOS attacks nor are they immune to total failure. I’ve written extensively of application layer hacking capable of directly bypassing firewalls thus raising the need for advanced behavioral analytics to help to identify breaches. Multi-vendor diversity can also help to lessen the risk of cyber attack.

3.1 ANYTHING-AS-A-SERVICE

Cloud technology has given rise to the XaaS model – what began as Software-as-a-Service has since evolved into Anything-as-a-service (Xmas). XaaS allows businesses to meet their customers’ demands without the hassle of needing to have expertise in every aspect of running a modern business. Today you can, for example, outsource your database to a database-as-a-service vendor or your customer support to a customer-support-as-a-service provider. Less savory examples of XaaS even include malware-as-a-service and DDOS-as-a-service. XaaS truly enables businesses to focus and provide value to their customers.

Recent M&A deals show signs of what I like to call “diagonalization” – a strategy whereby a company acquires or merges with another relatively unrelated company. This approach allows the two firms to form a conglomerate which often leads to new product extensions and offerings. A prime example is the recent Amazon/Whole Foods acquisition for a proposed $13.7 billion. I believe this acquisition has the potential to bring about one of the most sophisticated and complex logistical solutions on the planet. Pending FAA regulations, drone delivery services may one day allow for near-instantaneous delivery at the click of a button. I predict vertical and horizontal M&As will account for approximately 20% of deals moving forward while the vast remainder will be of a diagonal nature.  

Evolution of Datacenter

Modern businesses rely on the ability to easily integrate new products and services into their platform. It is important that today’s CxOs learn and understand the trajectory of cloud technology and learn how to apply it to their business. The image above illustrates my vision of cloud evolution. In general, I believe a well-designed cloud-based architecture that is capable of flawless delivery must include five major processes.

1.  Visualize – This was done back in the early 2000s and forms the basis of a modern cloud-based data center.

2.  Consolidate – Using the Concepts of One and Zero, data centers, networks, and systems both modern and legacy must consolidate to reduce complexity and form cohesion.

3.  Integrate – This is where policy-based management will be critical orchestration of not only cloud systems but also operational support systems (OSS) and business support systems (BSS)

4.  Automate – Work closely with vendors to deliver fully automated cloud-based systems that deliver best-in-class unit cost, capability, cycle time, and culture to provide the best possible service for customers worldwide.

5.  Federate – In essence, this is the culmination of each of the previous steps which result in an optimization of public and private cloud systems.

4.0 CONCLUSION

These five steps describe the evolution of the modern cloud-based data center. It is crucial that the modern-day multinational corporations understand and appreciate the current state of the cloud as well as its direct influence on the bottom line. The cloud is an amazing technology that has the potential to launch any business to the next level. But it must be implemented with great care and thoughtful design. It is exciting to witness cloud technologies become ever more pervasive and I look forward to seeing it continue to grow well into the 21st century.

The Harvard Innovation Lab

Made in Boston @

The Harvard Innovation Lab

350

Matching Providers

Matching providers 2
comments powered by Disqus.