What is Energy as a Service (EaaS)?

Energy as a Service is foundation of Sustainable Transformations at scale.
Source: VueLogix 2020

Background and Brief History

Simply put, it is perhaps our best tool in managing climate change. Traditionally energy usage has not been actively managed using contemporary IT tooling. We use the energy, we pay for it and rarely even think about what else can be done. This is a very antiquated approach, much like how we generate energy. A new approach is needed and the IT industry has the technology and ability to quickly innovate and transform how we live, play and work.

While still an emerging technology, Energy-as-a-Service (EaaS) follows a Software-as-a-Service (SaaS) model, with energy data and energy consuming operations being the focus.  Much of the framework we took for the industries first EaaS offer, Cisco EnergyWise we modeled on the Performance Contracting marketplace.

First, let’s make a clear distinction between Internet Protocol (IP) based energy management, versus the highly manual, performance contracting approach. The former uses automated discovery tools, asset management, custom queries, data marts on-demand and ultimately the ability to spin down or completely turn off any device with an IP address. Performance contracting by contrast, is based primarily on a visual site inspection, taking manual power readings from circuit panels, reviewing billing records, auditing operations and so on. The later also has little to no market in the lab and data center space. This is simply because an IP energy-based approach provides a scalpel, whereas performance contracting is more of a club.

How I arrived at the concept for Cisco EnergyWise, the worlds first Energy-as-a-Service offer might help to illustrate this newer approach. One day at Cisco, working in the Data Center Networking team, the call came out over email asking if we had any “treehuggers” on staff. The tone of the mail was dismissive if not derisive. I have an environmental sciences degree, had 6 years of data center electrical engineering experience and love to hug trees. So I put my hand up and became Cisco’s first chief sustainability officer (or at least a version of it; more like a sustainable solutions officer).

In determining what the IT industry (and Cisco therein) could do to help manage environmental impacts, it was clear the network played a massive role. In exploring what we could do as the worlds largest network vendor, I focused in on energy and the Greenhouse Gasses (GhG) associated. If you look at whatever device you are reading this on, focus in on the power or battery icon on your devices desktop/home view. Click on that icon and you should see an interface come up with any number of power management features. When I did that on my IBM Thinkpad and later on my MacBook Pro, an obvious question presented itself; why haven’t we simply networked these (local) power management features so a central, network admin could do that at enterprise scale. Just like desktop shutdown software of years past (1E, Verdiem, JouleX) but covering any and all devices on a given network.

With this epiphany in hand, I sought out Cisco Central Engineering, found my development partners and began the long process of developing a new product. It took less than a year to release the core EnergyWise offer but it didn’t become a true EaaS offer until we acquired JouleX to provide a best in class UI/UX and asset management framework. The full offer took 2 years to bring to market, about 1/3 of that time spent deploying EnergyWise internally at Cisco across buildings, labs and eventually data centers. In January of 2009 we launched EnergyWise and in my opinion marked the beginning of the “Green IT 1.0” era (~2008-2012).

Who is the User?

Just like any other SaaS offer, we sought to make EnergyWise an end-to-end solution. For the first time, we had an application that was shared across facilities and IT departments. The mismatch in systems and priorities between IT and Facilities departments is significant. Developing a tool that provides value to both was difficult but we pulled it together. We did this by focusing in on how energy operations were managed to identify the pain and inefficiencies inherent to the current state.

This led us to the potential users of an EaaS system. It was clear we needed to account for both IT and Facilities (building) managers. We found that Facilities pays the power bill and IT is often the largest user of that power. At the time Cisco was managing their energy usage the same way most still do, crudely (at least from an IT perspective). There was no active monitoring, consistent measurement or automated management in place. Facilities would simply bill IT quarterly/annually for the power it used based on a Watts/square foot ratio that was rarely re-assessed.

User Priorities

As we developed the technology to be more and more capable, automated and scalable; we found another challenge. It was clear very early on that facilities and IT departments rarely interact and jointly plan combined operations. The clear exception we found was in supporting purely data center customers (Google, AWS, Equinix, etc.). When a user has their data center as the primary service, their teams are far more integrated and efficient. There is a clear profit driver to be so.

For the rest and majority of our customers we found the following, apparently conflicting priorities between the 3 persona groups involved in an EaaS deployment.

IT Users

Focused on availability, uptime and resilience for critical systems only.  Expected to drive cost efficiencies using contemporary technologies but not at the expense of overall systems availability.  No direct incentive or pressure to reduce energy costs.

Facilities Users

Focused on availability, uptime, resilience and life safety services.  Not expected to drive cost efficiencies using technologies at comparable level to IT.  Top priority is human life and safety, critical systems availability and overall employee satisfaction with supporting facilities.  Pressured to reduce energy costs but have few options to do so that don’t involve changes to IT energy usage.

Executive Sponsors

Internal leadership ultimately needs to drive new, transformational models in general.  Energy is no exception and is net-new when approached in an ‘as a Service’ model.  Executive priorities are well served with this approach as it covers brand value (green), innovation (new internal green solutions and possibly new to market) and cost reductions.  The biggest challenge will be in coordinating priorities across facilities and IT teams.

Balancing Stakeholder Priorities

We found the most effective way to serve these diverse stakeholders was to make everything visible down to the user level. I cannot overstate the importance of making things visible that have been hard to see before. In doing that with energy we took Cisco’s internal approach from a pooled billing system where inefficiencies were hidden across a single, massive power bill to a very detailed view of what divisions within IT were using and which users were operating their assets most efficiently.

Just like the benefits your hear in Big Data and Data Science marketing; we provided targeted analytics packages, per-designed for specific user roles. These new capabilities provided a completely new capability to both facilities and IT teams that had significant monetary and environmental benefits. Furthermore, it took time and pain away from facilities teams by eliminating a number of manual steps and it gave IT teams the assurance they were only paying for the power they were using for their operations.

Operational Transformation

Making data visible in an intuitive model does a lot to change behavior but in today’s busy work-life, it’s not enough to just show the problem. We anticipated that users would want to take action if we showed how good or bad people were at the individual level. This led us to first established digital signage, locally outside all energy intensive environments (labs, DC’s, EBC’s). We ran “wall of fame” and “wall of shame” signage for 2 quarters while we were developing control policies. During that time, we had dozens of lab and data center managers approaching our team for the control policies we were already working on.

Example of an Energy Management Dashboard
Source: Talisen 2019

In addition to a trial, monitoring only period, we approached Cisco Workplace Resources (Facilities Department) leadership to request they assess the new program. Our goal was to familiarize non-IT users on the new monitoring/reporting system we deployed. Our value proposition to the facilities department was simple; you are paying more than you need to for IT’s energy usage. We can ensure you pay less and correspondingly reduce Cisco’s overall GhG footprint. They agreed to review and if comfortable with the new system, consider using it to redesign their current billing system.

With the IT teams our value prop was slightly more complex but equally valuable. Within the IT teams, we segmented out our approach to 3 different infrastructure domains that represent 3 different operations:

Office Space

Low criticality, predictable utilization, de-centralized.  Infra includes; phones, cameras, desktops, laptops, phones, tablets, WAP’s, switches, routers, printers, lighting, HVAC and specialized systems (video conf, badge security, etc.).

Labs

Mixed to low criticality, predictable utilization, semi-centralized.  Infra includes a wide range of devices and equipment.  Labs most often reflect a data center environment with similar equipment and support systems.

Data Centers

Mission critical, unpredictable utilization, centralized.  Infra includes servers, switches, routers, appliances, storage arrays, tape libraries and even main frames.  These environments should be the last focus area.  Starting with office space and labs will mitigate risk in eventually moving to production environments like data centers.

While we used the same EaaS approach for all 3 environments, we configured it to provide pre-designed management models for each. From offices to data centers, there are very different operational priorities. We started with the least critical infra (office space desktops, IP phones) and gradated to data centers. This stage of EnergyWise development moved past just monitoring and focused in on control policies. These control policies now gave network managers the ability to centrally administer the power usage of all devices on a given IP network. For office space and many labs, shutdown policies were fine but for any remotely critical, production related environment we needed to develop “spin down” policies to avoid any performance degradation.

Energy as a Service Deployed in Cisco Production Data Center, 2011

Why Transform?

When you look at your current organizations global electricity bill what do you see? Do you approach it with a “cost of doing business” mindset or is your answer “there really isn’t much I can do about it”? If you do, you’re like most. What you probably don’t realize is that in doing this at Cisco, we saved $47M USD (internally; 2011-2016) and delivered over $1B USD (2008-2016) to Cisco’s bottom line (externally; EnergyWise, SmartGrid, JouleX, Kinetic).

There are clear, significant monetary and environmental drivers for an EaaS offer. However, I think we were very early to market with EnergyWise. In addition, Cisco went through significant upheaval and transformation during that time. A focus on innovation seemed to leave Cisco somewhere around 2014 and doesn’t seem to have returned since (at least on Green IT).

The State of EaaS Today

Well, there’s good news and bad news. Let’s start with the good. IT as an industry started to become aware of environmental issues somewhere around 2008. I look at this time (~2008-2012) as a Green IT 1.0. There was a lot of attention, investment and even competition amongst tech giants to be the greenest player. I believe the market dried up due to a lack of public attention on the issue. Perhaps driven by more perceptibly acute crisis’s (war in Syria, global rise of authoritarianism) and maybe a lack of production ready solutions but either way it dried up.

The pandemic seems to have driven attention back to the bigger, global issues that effect us all. It appears we entering into a Green IT 2.0 phase. Unfortunately it appears we are not learning from Green 1.0 as of yet. Cisco is still absent, IBM, HP, Google, Microsoft, Dell, EMC, Hitachi, Apple and just about any other large tech player are as well. All of these companies produce wonderful annual reports that promote all their efforts to be more green internally. While I sincerely applaud these efforts, they are all Green 1.0 at best. It’s not enough to hand wave anymore if you are a large IT vendor. IT vendor’s are supposed to innovate and build solutions to the hardest problems. Where are our IT vendors in providing solutions to the most important challenges humanity faces? While the IT industry spends valuable brain power on driver-less techs and robot vacuums, the climate burns around us. When will our IT giants step up to the challenge and start providing sustainable solutions to others. Why aren’t these companies providing others with a blueprint to be as good as they are?

“All the large tech players; Apple, Google, AWS, HP, IBM, Cisco, Dell, EMC, market to their shareholders the many sustainability measures they’ve implemented in house.  To date, none of them have decided to take those learning’s external.  Since Cisco exited in 2017, no other technology player has been serious about providing solutions to help others be more sustainable.  The market is ready for new solutions, we need “earth innovators” in IT to take us to the next level.”

What is needed to bring back An IP-Based EaaS?

While I’m hopeful the larger players will start to innovate for the Earth, the only one (of any size) I see doing anything external is Salesforce. They released Sustainability Cloud around the same time as the onset of the pandemic (worst possible timing perhaps). It is likely this space is now well understood enough that a smaller player is well positioned to disrupt this space.

A small player may not have the brand and deployed base of a larger player but is less likely to try and reinvent the wheel. Larger players are often stuck in legacy thinking, internal politics and a “not invented here” attitude which all work against innovation. Whomever looks to take on the next level of EaaS development, they will still need to recreate much of what we did with EnergyWise. If you have an interest in looking into this space further, I’d suggest starting where we started with these basic capabilities:

Energy Consuming Devices (IP)

The vast majority of devices today that communicate via IP already provide energy/telemetry data out of the box. Some basic scripting and data gathering schemas can aggregate and report on energy data today.

Energy Consuming Devices (non-IP)

In addition to IP networks, there are massive Supervisory Control and Data Acquisition (SCADA) networks deployed globally (utility and manufacturing lines for example). These networks can provide energy data but it requires some additional steps to land. There is a lot one can do without interfacing with SCADA networks but a true EaaS system by providing flexible edge connectivity.

Energy Discovery Protocol

There are far too many physical devices on most enterprise networks to interrogate manually. Therefore an automated discovery protocol is required to establish initial benchmarks for energy usage.

Energy Monitoring Connector

Once discovery is completed, a capability to now monitor to the asset discovered is required. This provides an ongoing data feed for energy usage.

Energy Domain Builder

As assets are discovered and monitored, management of the assets and association to their managers is required. A simple domain builder (similar to network management models) is required.

Energy Control Policy Library

Continuing past benchmarking and monitoring, a step is required to action an energy reduction. This can be accomplished by developing control states that either shut or spin down certain processes within a given asset (server, switch, array, etc.). A library of policies, designated against different infra classes will be needed to drive direct savings from an EaaS system.

Energy Manager Interface

Net-new energy managers are required to run an EaaS system. At Cisco we found that this role could be performed as a part time focus for a network manager, working with a facilities lead counterpart. We built a well designed managers dashboard to support these users.

Energy User Interface

In addition to the management interface, a consumer view should be developed. Reporting directly to the users of energy within an organization has major behavioral implications. Whether you take a fame or shame approach, making it visible will drive change.

Energy Dashboards

A final, external dashboard should also be considered. While we were not able to get this done in time at Cisco, an external public feed on how good your organization is doing (in real time) against its environmental targets is a significant differentiator.

Who Should Look at This?

I think Cisco clearly made a mistake in letting EnergyWise go and stepping away from the market they to a large extent, founded. I think this is largely due to the lack of innovative thinking at senior levels. That said, it’s wide open now and most of what we did with EnergyWise could be recreated quickly by Cisco’s usual competitors (Palo Alto, F5, Huawei, Arista, etc.).

However, I think this may be more of a start up play now, particularly for those players who see the full potential of Software Defined Networking (SDN). The network is clearly in the best architectural position to make energy digitally manageable but really anyone who can build a switch/serve/compute device can do it.

For the majority of the clients I speak to on this topic; they are all waiting for a solution in this space. While many of them like Salesforce’s new offer, they clearly see that this latest tool doesn’t provide any of the energy as a service features that died with EnergyWise. SF does have the resources to do it but I suspect sustainability is just too far from their core to get much attention internally.

Would you Deploy EaaS for a 40% Reduction in Electricity Costs?

I mentioned earlier that at Cisco we saved almost $10M/year for 5 years. That netted out to about a 40% reduction (averaged YoY) in total electrical spend related to IT electricity cost allocations. This is perhaps the best kept secret in the history of Green IT. Almost everyone in IT has gone the renewables and/or carbon offsets route, but few if any other than Cisco have implemented a purpose built energy and environmental management system. This was done at a relatively low cost with a team of less than 8 and less than 2 years to bring to market.

If you are an IT executive, please recognize you will have to pursue these savings in tandem with your facilities counterparts.  The technology, deployment and book have won awards. The savings are real and many of the project members are still working in the same space. However, it would seem that the word is not out and the approach is not well understood today. That is why I wanted to write this article to get your take. If you’re an executive, what if anything new are you willing to do to innovate against climate change? If yes, how did you prioritize and where did you land on it?

Related Posts

Leave a comment