Does (which) Cloud Matter?!

In May 2003, Nicholas Carr’s blog “IT doesn’t Matter” and related book “Does IT Matter” created quite a stir in both business and technology circles. I was transitioning careers, from engineering at HP into consulting at McKinsey then. It was a great topic for business school interviews. Thirteen years later, I think it is worth asking a similar question, “Does (which) Cloud Matter?!”

As Enterprise CIOs evaluate architectures for their cloud infrastructure (public, private, hybrid or multi) and more broadly their IT infrastructure, they should consider the following important points. There are also some implications on opportunities that arise from these considerations.

1. Cloud (IaaS) will become a commodity:

With Amazon, Microsoft and Google leading the charge on public IaaS Clouds, along with a host of other cloud providers following (including IBM SoftLayer, and telcos), Cloud computing will become a commodity. Commodity means that it will be hard to distinguish one cloud from the other on features and functionality like a utility. Cloud providers are innovating at a rapid pace and introducing features at a rapid clip. One of my engineering friends accurately put it, “Anything I can think of as a feature, these guys already have it announced and launched and are doing at a pace that is very hard to keep up”

Commodity also means that IaaS cloud service vendors will need to have a differentiated cost structure and customer experience to win. (think Auto Insurance, home lending). It doesn’t mean that prices will drop indiscriminately or that it will not be a profitable business (think telecom service providers and oil and gas companies).

It also means that most enterprises will not be able to match the scale, operational discipline and technology automation of vendors. Enterprises, will therefore have a choice of basic IaaS services and may need to manage across providers.

2. IT Operations (for most enterprises) will become more like supply chain management

IT Operations and IT Infrastructure in most enterprises should eventually go the way of Supply Chain and Manufacturing. This is a good thing and a process of evolution of the important function IT performs. Manufacturing as a function changed dramatically in the last few decades. Supply Chain excellence became more important than manufacturing excellence. The best run supply chain organizations (e.g., Cisco, Apple) focus on where they can innovate – vendor management, forecasting, design for cost and quality.

In the Cloud context, large cloud providers will even the playing field for most enterprises on access to IT Infrastructure. Which cloud you pick will be less important than alignment of operational SLAs with business priorities – cost, performance, security, and availability. For example, speed may be more important than security when there is no confidential data involved. Therefore, Enterprises need to develop operational excellence similar to supply chain to fulfil their mission critical role in the organization, especially in the context of public IaaS and SaaS offerings.

If “which IaaS cloud” doesn’t matter, then where does the Enterprise customer innovate? And what does it mean for start-ups and investors in this space?

1. Focus on what you do with the cloud – a cloud architect is critical.

The Fabric hosted a panel recently on hybrid cloud trends and opportunities. The panelists (two vendors and two customers) unanimously agreed that cloud applications that create differentiation is the highest order bit. You need someone who can understand applications and business to drive your cloud architecture.

In this context, the cloud architect becomes one of the most important hires in your organization. She/He must understand all the technology innovations, its complexities and should map it to the business needs. This is a trend Cisco surfaced a couple of years ago. Cisco’s sales teams realized the importance of having a peer to the customer’s architect and continue to hire such talent .

2. Strive for non-intrusive supervision – It is necessary but difficult

This was the second point that came up in the panel. One of the biggest impediments for adoption of cloud – public, private or hybrid – is the perception of lack of compliance, security and availability. Reducing these risk factors has historically been associated with the CIO. The challenge with the current technologies is not that these are lacking but there is lack of visibility on what is available.

The answer to this problem is not to put fences and limit the use of Cloud technologies but to have supervision that is non-intrusive. I liken this to parenting. The first approach is similar to locking your kids in the house to prevent a scrapped knee. If you did this, they will never know how to ride a bike, come down a slide or roller blade. The latter is similar to teaching your kids to bike in a park before letting them ride on the streets with a helmet.

3. Listen to the developers

One of the history lessons of technology (e.g., Microsoft in the 80’s, Sun’s Java in late 90’s, Android/Apple in Mobile and now by Amazon in the cloud) is that the developers define the platform. While a few niche enterprise and developer requirements may lead to a mix of public/hybrid/multi cloud environments for enterprises, developers should be the dominant influencers of the cloud architecture and hence the corresponding choice of the IaaS Cloud provider.

Developers have a tough job today already (understanding the business, managing feature creep and changing deadlines, adapting to limited talent and resources). Therefore, operations should not add complexity or friction to development and deployment. Developers will, therefore, continue to gravitate to the easiest and most agile development environments. This will put more pressure on Operations (and DevOps) to get the deployment cycle simpler and faster. The logical solution response will be to develop and deploy on the same system.

4. Anchor on Data Management – it will become critical.

The last piece of the puzzle is data management. Enterprises will continue to contend with data sprawl. Cloud will add to this. Enterprises will have data sets that are spread across multiple clouds (public, private cloud providers, embedded in third party in SaaS providers and even mainframes).

How will enterprises handle this? Enterprises will need to maintain consistency of the “system of record” (e.g., banking transactions), simplicity for “system of integration” (e.g., customer sales data) and velocity for “system of engagement” (e.g., customer trends). They will experiment with many approaches (e.g., tools, cloud information bus) before adopting one that is right for their needs

In summary, while Cloud is a large macro trend, Enterprise customers focusing on using IaaS clouds and their capabilities effectively stand to win. Those who fixate on which specific Cloud technology to use or fail to understand and explore the agility and advantages will be consigned to be laggards in their respective markets.

What is your experience? If you are an enterprise, what is your biggest hurdle in this journey? What new technologies and innovations will enable this future and accelerate this will be the multi-million-dollar question.  At The Fabric, we are eager to hear your thoughts and needs.

Rajesh is an entrepreneurial executive who prides himself on driving ideas to execution as a core member of any management team. He is a broad experience as an engineering leader, CxO advisor and leader of DC strategy of a multi-billion business. He has co-authored publications and thought pieces on infrastructure and storage.

by Rajesh Srinivasaraghavan, VP Business Development at The Fabric