Zero Lead-Time Computer “Hardware”

By Chris De Vere Let me share an idea with you: The manufacturers of all products have the same problem regarding delivery lead-time: How can I get my products to my customers when they want them? The traditional approach is to use a two tier distribution model: a wholesaler who buys in bulk from the […]]> By Chris De Vere

Let me share an idea with you: The manufacturers of all products have the same problem regarding delivery lead-time: How can I get my products to my customers when they want them? The traditional approach is to use a two tier distribution model: a wholesaler who buys in bulk from the manufacturer; and a retailer who uses his understanding of his customer to stock the items that he believes his customer will purchase.

There are three problems that I see with this approach: the first is the issue of multiplying probabilities; the second is that the more combinations on offer the lower the probabilities are likely to be, and the third is that the more perishable the goods the shorter the shelf life. Let me explain.

Supposing that a distributor knows what to purchase for his reseller customers 9 times out of 10; and let us assume that the resellers know what to purchase from their distributor for their end user customers 9 times out of ten (0.9 x 9.0 = 0.81). This means that the reseller will actually have what the end customer really wants 80% of the time.The more options a product has, the more likely that the probability will be lower.Let us say that with more options, the distributor knows what to purchase for his reseller customers three quarters of the time; and likewise we assume that the resellers know what to purchase from their distributor for their end user customers three quarters of the time (0.75 x 0.75 = 0.56). Now the reseller will actually have what the end customer wants in just over half the time!Only having what your customer wants in half the time is not a recipe for success.This problem of options is compounded if the product is has a short shelf life, like fresh food, fashion or consumer electronics.

We have seen retailers come up with some innovative solutions: on-line grocery shopping with delivery to the home, the customer ordering much closer to what he actually wants. We have seen Amazon.com make the news about experimenting with drones to deliver goods direct to the customer to reduce lead-time.

Let us look at computer hardware, and in particular the world of converged infrastructure. Here we have the problem of a lot of options and limited shelf life. As a result, these products are often built to order.The market leader in this space has a 60-day lead-time. Another supplier has a relatively short 20-day lead-time. All offer some sort of scheme where the customer can include some sort of “buffer stock” to eliminate lead-times in the future. Typically, these are blades that are not charged for until they are used. One company offers a buffer stock with a “rental,” but rent this for a month at the exorbitant rate charged and you will have paid the original purchase price. So, you can see that all these buffer stock and rental plans are really variations on the same theme: clever financing.

How can an IT organization overcome the problem of computer hardware lead-time without clever financing?Here is a way.You ask your hardware supplier to deliver VMware’s vCloud Hybrid Service as well in the interim. Almost straight away you will have computing resource available to use while you order exactly what you want from your supplier and wait for delivery.This is what Interop had to say when awarding vCloud Hybrid Service the “Best of Interop” award:

“A customer doesn’t just initiate a workload under the VMware hypervisor in the cloud; he initiates one, when desired, that’s identical to the one running in his data center. The network characteristics are the same. The policies on the firewall and other security measures are the same.

To the system administrator in the enterprise, the new workload may be in the cloud, but it’s managed from a familiar vCenter management console and it looks and acts just like those across the data center floor. In particular, VMware’s efforts to virtualize the network and assign it as a flexible resource to a virtual machine at the moment of creation means a system administrator can stretch layer 2 and/or layer 3 networks seamlessly from the data center to vCloud Hybrid Service without the need for manual configuration changes.

In the past, moving the workload has been possible into a public cloud, but that necessitated using different tools and monitoring systems and setting up networking on the cloud provider’s terms. VMware is making it possible to extend the server, networking and storage envelope inside the data center out into the public cloud, making it an extension of the enterprise.” – Charles Babcock

So, the answer to solving the lead-time problem is to order vCloud Hybrid Service when you order more hardware. Logically extend your IT environment using your Security Policies and Network Parameters and later copy your workload back in-house when the hardware is delivered and installed.If you are running a VMware environment, no conversion of the workload is required as would be the case with other clouds.

The question is, what would it mean to you if you could start using more IT resource immediately rather than waiting one or two months? And as Nicholas Negreponte, the guru at the MIT Media Labs, said: “Move bits, not atoms.”

Chris De Vere is vCloud Business Development Director within VMware EMEA’s vCloud Hybrid Services team. With VMware since 2011, Chris worked to establish vCloud as the leading Cloud platform for Service Providers across Europe Middle East and Africa, and now works to bring VMware’s own vCloud Hybrid Services operation up to speed across Europe, Middle East and Africa. Prior to joining VMware’s vCloud organization, Chris worked at IBM in IBM’s fledgling “Cloudburst” team, EMC, Silicon Graphics and Verity.

via VMware Blogs http://bit.ly/R3PAiN