I had an update from the VCE organization this week and even though I’ve been covering VCE and Vblocks since their inception, there was one thing I clearly didn’t understand and that was: This solution is an appliance. This is why it can go from bid acceptance to production in as little as 40 days; this is also why it is generally cheaper, more agile and more scalable than more traditional solutions. In effect, this is the system that Henry Ford would build and it represents as big a change as what happened in the auto industry under his guidance to the large-scale computing market. It also turns primary partners EMC and Cisco into parts suppliers.
I’ve been focused on the uniqueness of the partnership and missed just how unique this Vblock offering is.
Custom On-Site = Complex, Unreliable, Costly
Systems at this scale, and we are talking at the extreme side of the IT market, tend to be highly customized and assembled and tested on site. This is largely why the time it takes from when the system components are first shipped to a customer site and when the system is fully configured, out of test and in production, can be measured in years. This long cycle has a number of inherent problems that almost assure that the system won’t be completed on time and within budget.
Much like it would be if you were to build your own unique car from scratch as the system is built, assumptions are tested and found wanting, which increases the time to completion during this time. Goals and software changes force additional changes and as the system begins to grow differently than first intended, initial assumptions become obsolete, forcing changes to everything from how the system is provisioned to how it is cooled and powered.
The end result is that processes used to build a custom system on site are far more complex, far less reliable and far more expensive. Building on site not only leads to a more complex and unique system initially, but the time it takes to complete the project results in changes that increase this complexity and cost.
Appliance Approach: Faster, Better, Cheaper
The VCE Vblock Appliance approach is far more efficient. The specification has to be tightly defined before the project starts. The product is completed and shipped in days so that there is both little opportunity for in-process changes to be needed, and, because this is an appliance, those changes are contemplated as post-completion modification and not an in-process change. This last is important because the change is applied to a known working system and not during the manufacturing process and before the planned system is tested for quality and compliance.
You see, it is far easier, and better, to modify a known working system than to mess with a system that hasn’t yet been assured. When you modify something in process, you are not only likely to break it, but you are far less likely to understand what caused the break. If the system is assured and you make a change and it breaks, you know the change was likely the cause.
Wrapping Up: Lessons Learned
This VCE meeting reminded me of the importance of the appliance approach to both providing a more reliable solution — just like the first Fords were generally more reliable than other cars of the time — but also one that is potentially far cheaper to install and run.
My only concern from the meeting was the sense that VCE was resisting the use of EMC’s leading efforts in customer loyalty and satisfaction, something I’d been assured would naturally happen. I think this showcases that even the perfect partnership may have a few issues that still need to be addressed.