Amalgam Insights specializes in Technology Consumption Management with a focus on how companies can align IT purchases and subscriptions to the relevant consumption models. This can and should include options for enterprises to purchase not only on a per-server or per-core basis, but also on a duration, user, resource, usage unit, feature, or baseline capacity basis. In this light, AI was especially interested in evaluating HPE’s Flexible Capacity billing capabilities across storage, computing, and networking as well as HPE’s new launch of Project New Stack.
HPE Flexible Capacity: Cloud Consumption for Hybrid IT Environments
HPE’s Flexible Capacity is a usage-based billing service for HPE infrastructure, including server, storage, networking, software, and services installed on an on-premises basis. This service is focused on customers expecting to expand IT infrastructure over the next three years while seeking a pay-per-use model.
Prior to purchase, companies should have an idea of
• core server, storage, and networking capacity. This provides a baseline for costs, since HPE will typically charge the customer at 100% of forecast in the Capacity Basic Model or a minimum of 80% of the forecasted capacity for the Capacity Premium Model.
• expected buffer capacity in excess of core capacity that should be deployed with the initial deployment. If this buffer is undefined, HPE assumes a 10% buffer capacity that is provided, but not billed unless the buffer resources are actually used.
• units of capacity to be billed, which can be based on servers, cores, GB-hours, ports, or licenses based on the level of granularity that is appropriate for the client.
Based on this, HPE will charge based on the capacity, usage, or resources actually used based on a minimum commitment. However, past that minimum commitment, resources can be dialed up or dialed down based on peak needs, business changes, or other business events that affect IT demand. If a company approaches its initial capacity buffer, HPE can provide additional resources based on the company’s new capacity forecast.
In addition, the pay-per-use model will also include install, implementation, maintenance, support, and upgrade resources to provide a cloud-like cost model. HPE provides Datacenter Care support and maintenance resources for break/fix, on-site support, and additional support based on the client’s contract
As HPE develops newer storage, server, or networking products, HPE also provides an optional refresh sourcing option for products that are typically 48+ months old.
Flexible Capacity also includes the potential to purchase Microsoft Azure Public Cloud within the SOW. In this model, the Microsoft Azure services will be handled on HPE’s paper, including contract, invoice, support, and the Flexible Capability Portal based on HPE’s Project New Stack which includes functionality from Cloud Cruiser, a cloud usage and expense management solution previously licensed by HPE and acquired by HPE in January 2017.
HPE also provides monthly overviews of capacity usage as a basis for the capacity fees charged. This detail includes:
• Inventory of servers, storage, and networking resources
• Device-specific and monthly aggregate usage data
• Content, terms, and start/end dates for the system schedule
• The applicable unit rate
• Problem/incident status report, and
• Change request reports
HPE Launches Project New Stack
At Discover 2017, HPE announced a preview of what was called Project New Stack https://www.hpe.com/us/en/solutions/cloud/hybrid-it-management.html. Based on what HPE showed, AI believes that this will be an important management plane of control for managing composable infrastructure and for managing hybrid IT costs. For HPE hybrid IT management purposes, AI believes that IT managers who are either HPE customers or considering HPE purchases (including Nimble and Aruba) should be aware of this emerging capability.
As shown at Discover, Project New Stack is designed to manage three key tasks associated with Hybrid IT management: managing applications, operating environments, and infrastructure resources associated with a hybrid and composable IT environment. This management plane is extremely important because even the most basic usage-based services can quickly blow up from an expense perspective.
Consider the world of telecom and mobility as a precursor for IT management. In these sub-sectors of IT, telecom and cell phones were traditionally treated as operational costs that could be managed by an IT manager, operations manager, or office manager and these purchases often occurred outside of the traditional IT purchasing process. And then suddenly, enterprises discovered that the upside cost potential of an unmanaged circuit could run into thousands of dollars spent outside of IT governance. Even worse, the cost potential of a single unmanaged cell phone was not $200-$300 a month, but potentially $25,000 or more as overseas roaming data charges got out of control.
Cloud has a similar problem, both in managing the cost of specific instances and the ongoing governance. Without management, it is easy for a budgeted $10,000 cloud instance to turn into a usage-driven $50,000 cost. Even worse, if this cloud instance was intended to be a short-term investment and gets ignored over time, it can turn into a recurring $50,000 per month cost. Now, multiply this cost by the number of infrastructure projects, independent developers, and locations in your organization. The potential for ungoverned cost overruns is massive in a world where developers can potentially set up unlimited computing costs with little to no oversight.
In addition, if companies seek to conduct comparison shopping, it is difficult to conduct an “apples-to-apples” direct comparison across clouds at this point. By looking at assets across multiple platforms in a single management pane, enterprises are better empowered to make these direct comparisons.
AI believes that Project New Stack has additional potential to support increased financial and IT governance through the development of contract management, dispute management, a shared glossary or ontology of cloud elements, integration between orchestration and financial management, and integration across IT finance, corporate finance, HR, and procurement systems. Based on the initial preview, AI looks forward to seeing how Project New Stack continues to develop as a potential bridge between the CIO and CFO offices and competes with independent vendors such as Cloudyn, CloudHealth, RightScale, and IBM Cloud Asset Management.
AI Recommendations to IT Technology Managers
AI believes that IT managers should take the following concerns into account as they consider HPE Flexible Capacity.
1) Negotiate custom support terms carefully, as these terms can be unique to each statement of work. This includes comparing HPE’s standard support to the expectations of internal Service Level Agreements and public-facing service guarantees. Although there are standard support terms that HPE provides, enterprise IT organizations must ensure that capacity is added on or removed while retaining enterprise support standards.
2) Consider volatility and risk profile of the company in looking at Basic vs. Premium models. HPE’s Basic Model assumes a baseline of 100% of server capacity as a starting point for resource payment while the Premium model assumes a baseline of 80% of server capacity. Companies with cyclical or volatile revenue expectations should purchase accordingly based on a risk-portfolio based estimate of their resources over the next year.
3) Define your Buffer Capacity ahead of time, which determines the excess resources your organization can pull from. Is HPE’s assumption of a 10% buffer sufficient for your organization? If not, which business events would increase that buffer capacity needed to be effective?
4) Be proactive in defining the change management process for increased buffer capacity replenishment. For Flexible Capacity to be similar to the cloud from a consumption perspective, IT departments must do a good job of governing and projecting future workload needs that are either planned or predictable. This means tying developer workload estimates and requests with IT capacity. IT is increasingly tasked to be resource accountants as well as skilled engineers as the cliche of “running IT as a business” becomes a reality.
5) Understand the limits of Flexible Consumption. Currently, this approach is focused on HPE and Microsoft Azure Cloud Services. Companies using multiple vendors for IT infrastructure or seeking to unite infrastructure, mobility, and telecom spend under a single view must still conduct additional contract and sourcing aggregation to have truly composable layer for hybrid IT sourcing and procurement. Though Project New Stack will be able to provide a version of the truth for hybrid IT usage across clouds with cost center drill-down, the sourcing of non-Microsoft cloud services will still be based on paper outside of HPE at this time.