Nabugu -

HPE exec: GreenLake is to HPE as AWS is to Amazon

In this Q&A, HPE's Omer Asad discusses the vendor's transformation from an on-premises hardware provider to a platform company with its GreenLake portfolio.

Hewlett Packard Enterprise is all in on GreenLake, its take on infrastructure as a service. The hardware vendor wants to bring as much of a point-and-click, cloud-like experience to customer data centers as it can while continuing to provide a more traditional buying model.

At the recent HPE Discover conference in Las Vegas, the vendor emphasized its transformation from a hardware company into a platform company, where storage, compute and tools are all offered as a service. The transformation has been several years in the making, starting with a push to lease its hardware. Now, it's focused on providing a cloud-like consumption model to customers.

In this Q&A, Omer Asad, vice president and general manager of data infrastructure and SaaS platforms at HPE, talks about what GreenLake means to the company, the future of hardware and why customers care less about speeds and feeds than they do making sure they have the right infrastructure for their workloads.

What is GreenLake to HPE? How should customers think about this sprawling portfolio?

Omer AsadOmer Asad

Omer Asad: The easiest way to understand this is: What the AWS is to Amazon, what Azure is to Microsoft, that's what GreenLake is to HPE. We would love to sell most of the products as services. There is a lot of feedback from our customers saying, 'We want the infrastructure consumption to be simple.' If they want, they can continue to buy compute appliances from us, they can continue to buy storage appliances, they can buy a lot of networking appliances, Aruba-based appliances. But increasingly, customers want what they call a cloud operational model and a cloud consumption [model]. Meaning, I just want to buy a service and I want to pay per month or whatever the metering technology may be. Some of those services may accompany some hardware along with a compute service.

Is HPE's focus shifting away from hardware development to how it can better serve particular workloads?

Asad: In order to keep up with the throughput requirements of the world, we will continue to remain ahead in many cases of speeds and feeds. A platform organization is a big organization, it's a thumping organization, but how a customer consumes that platform is going to be as part of a service. We still have customers that like our block service but really want to understand what platform this is on. We open it up; we show them the NVMe fabric speeds and feeds, microsecond latencies, dedupe factors, everything. But what we're noticing is that a majority of the customers are just trying to purchase an outcome at the end. And that outcome needs to meet their ROI.

Talking about the speeds and feeds model or how much theoretical maximum performance measured in IOPS and bandwidth -- is that fading as consumption models change?

Asad: Customers don't care about that; all they care about is the RAID card. [For example,] they need a mission-critical service from HPE with 1.2 million IOPS, guaranteed latency, and want to know the price per gig, per month. That service will arrive on our hardware that has to be able to perform that service.

What customers are you targeting with GreenLake and what are they most attracted to?

Asad: Smaller customers are much more amenable to GreenLake as it can simplify their balance sheets. Larger customers have accounting practices, where they like to depreciate the hardware themselves. Those are the customers that want to buy Capex.

There are a few customers that are what we call 'in transition' with us because we ourselves [are in transition]. When we first came up with GreenLake, it was, 'Buy all our hardware, and we'll lease it to you.' There was no concept of cloud operational model to go along with the cloud consumption model. Now what GreenLake is, what we call the GreenLake platform, you log into it, just like you log into your console on, you log into console or GreenLake, you buy EC2, EBS, S3 services from us, you buy compute storage and networking services. When you buy these services, these services come with hardware software combined inside of that, and you start consuming.

Most of our customers that are still buying customized pieces of hardware and leasing it from HPE; those customers … still want to custom fit somewhere, which is OK. [With GreenLake,] it's a flexibility option that they desire. A majority of the customers like the AWS or hyperscaler style of consumption.

What about integrations of existing hardware with GreenLake? For instance, a customer uses Pure for storage in their data center. Can GreenLake easily integrate with that?

Asad: No, that's a challenge that we have. For example, we have partners that are part of the HP Complete Program, like Qumulo, Commvault, Veeam and other companies where we take the entire infrastructure and put it on a single bill for the customer … When you have Pure in the picture, then the customer is paying for two subscriptions. Because you can get Pure-as-a-Service, which is Pure Storage leased. The customer actually pays for the GreenLake compute service and Pure Storage service.

Let's say users had a different storage -- say a white box on their own. Could HPE GreenLake easily integrate with that?

Asad: We have the capability of incorporating and coming up with custom GreenLake contracts. We try to minimize this. In order to facilitate [integrations] properly, we have to populate and fill out our console continuously with newer services.

Storage as a service sounds good on paper, but customers are concerned since their storage capacity needs can outstrip all other needs. How is HPE handling this problem?

Asad: We don't charge penalties for overconsumption; there are grace periods. Because we're constantly monitoring and metering their storage, if a customer has a minimum commit, and then continues to go over it, we just keep sending our boxes out. Customers can adjust their contracts. HPE offers five-, three- and one-year terms.

Adam Armstrong is a TechTarget news writer covering file and block storage hardware, and private clouds. He previously worked at

Dig Deeper on Storage architecture and strategy

Disaster Recovery
Data Backup
Data Center
and ESG