Craig Nunes, vice president of marketing, 3PAR: 2010, Battle in the Clouds
January 2010 by Craig Nunes, vice president of marketing, 3PAR
With cloud computing and virtualisation becoming more mainstream by the day, 2010 will be the year in which infrastructure vendors selling a “complete stack” of virtualisation hardware, software, and services battle it out with those selling individual, “best-of-breed” products. At stake: who will win market leadership and customer dollars as customers move from hosting their own applications to running them over the cloud.
“Best-of-breed” vendors, as the name implies, stake their claim on doing one thing better than any other vendor while also making their offerings easy to integrate with other best-of-breed hardware or software products. With this approach, the pitch goes, the customer gets the best network switch, blade server, storage array, hypervisor and database available. The customer receives the best performance, reliability, and security at the lowest possible cost without being tied to any one vendor’s proprietary technology. Examples of best-of breed products in the virtual datacentre include network switches from Cisco, blade servers from HP, server virtualisation from VMware, database software from Oracle, and utility storage from 3PAR. The customers who are most capable of tying all these best-of-breed offerings together are cloud service providers, since integration is their core expertise and in many ways how they distinguish themselves from competitors. Their integration know-how and economies of scale, is also how they argue that they can provide better, or less expensive, services than companies trying to create their own “private clouds.”
On the other side of the ring, the “complete stack” people are betting that at least some customers prefer a fully integrated datacentre stack, even if certain components do not measure up to the best-of-class products. These vendors are offering networks, servers, storage and software pre-configured under one brand, ready to go. Such an approach ought to appeal to small-to-medium businesses (SMBs) or small-to-medium enterprises (SMEs) who often rely on the reseller channel for integration anyway. Examples here include EMC’s Virtual Computing Environment (VCE) initiative with Cisco and VMware, HP (with its acquisition of 3COM and EDS), and database giant Oracle (assuming its acquisition of Sun Microsystems goes through, giving it Sun’s server, storage and operating system technology). The largest, savviest cloud computing providers – companies such as Savvis, AT&T Global Hosting, and Terremark – would initially appear to be another prime market for these packaged offerings given the massive infrastructures they must deploy and support. However if these companies all choose the same “stack,” they’ll find it harder to differentiate themselves, or to gain a sustainable cost or performance advantage over their competitors.
“Why should I not bet on a vertical stack approach?” is a question I often get from customers, competitors, and industry analysts. At first glance buying a “vertical stack” of technology from a single vendor would seem to be a good bet for SMBs who lack the skills or staff to integrate such products themselves. But will it be as attractive to a cloud services provider? After all, these cloud providers are looking to sell cloud services to many of these same SMBs the vertical stack players are courting.
Cloud computing is creating new profit pools in the IT industry and cloud service providers are in a good position to claim their fair share. Using the same vertical stacks of technology makes it harder for these providers to differentiate themselves from each other and shifts the profit pool to the infrastructure providers. In fact it becomes a competitive necessity for a cloud provider to build the most competitive stack they can – better than its competitors and differentiated by their expert choice of best-of-breed options – if they want to win in this new market. It also should make an SMB wonder why they should buy services from a cloud provider who’s just running the same vertical technology stack they could buy themselves.
The “complete stack” vendors have every interest in protecting their positions of incumbency in today’s enterprise IT market. Their creation of vertical stacks is likely to be their weapon of choice as they fight to slow the adoption of cloud computing delivered by service providers. After all, service providers are more efficient and will, on the whole, buy less equipment to meet a specific business requirement when IT function is moved into the cloud. That is bad news for the incumbent infrastructure leaders, and why they are trying to push these new complete stacks to enterprises, and to the service providers with whom they are really competing. However the new generation of cloud computing providers is starting to recognise this threat to their unfolding business models and is becoming increasingly sceptical of the vertical stack vendors.
Assume that over the long term – ten years or more – enterprise IT will be increasingly delivered as a service, either by independent cloud service providers or from enterprises so expert in IT that they offer “private clouds” to their own users or even to other customers. Those who cannot handle the integration themselves will find that they have to choose between buying cloud services from an outside provider, or the infrastructure to do it in-house from a single-stack vendor. I predict this long-term trend towards the cloud will ultimately win since the economics and flexibility of a cloud “utility” can be widely demonstrated in businesses small and large today, ultimately minimising the value provided by single-stack vendors. I also believe that those single-stack strategies sold to shareholders on the basis of cloud adoption will fail to deliver the results promised, and will be a doomed attempt by industry executives to justify their merger and acquisition strategies as a way to reignite revenue growth.
Time will tell, but smart customers might want to take a close look at the technical strategies of today’s leading-edge cloud service providers to see whether these industry-leading players are choosing complete stacks or the best-of-breed approach. They’ll find many of these heavy hitters have even developed de-facto “reference architectures” using best-of-breed products to drive up flexibility while driving down cost and vendor lock-in. Examining the technical strategies of these big service providers, who live and die by the cost and quality of their service, might make a good playbook for those of you trying to choose the right path for your own datacentre.