Subscribe
Category
White Papers
Demo & Video
- Home » Blade Servers » Truth in Analysis: Is UCS Really a Money Saver?
Product Overview
Truth in Analysis: Is UCS Really a Money Saver?
Posted November,11,2010 by Dan Busby
Cisco recently posted a case study on deploying their UCS at a customer site, Slumberland.
Once again, I’d like to dissect the study and cut through the marketing Spin and FUD. I’ll approach this analysis the same way I did back in April and then June ’09 when Cisco also tried to Snow the market about its pricing.
So, Just the facts, ma’am:
In the Slumberland study, Cisco is trying to grab the responsibility for massive cost reductions of up to 75%! But, when I read the details and do the analysis, the lion's share of cost reductions are related to the consolidation of physical servers to virtual servers. The case study states that there are now four blades…to host 56 virtual servers. There is mention of the IT support costs for those 56 physical servers and how the savings dropped the IT support costs from about $1,500 per physical server to $80 dollars per virtual server. I wonder how Cisco could be responsible for that? Seems like this would be better as a Hyper-V case study…
Now don’t get me wrong, unified computing will certainly simplify data center management – Egenera has been using a similar conceptual architecture to UCS, and has thousands of installations that enjoy the benefits – with significant efficiencies that result in operational savings in the long run – But let’s be clear that majority of savings in this case study are not coming from the Cisco Unified Computing System.
If you want to look at Cisco’s contribution to savings, we need to compare the costs of the Cisco UCS hardware against the competition – HP, IBM , DELL, SUN, Fujitsu. These vendors have some incredible price points – a blade with dual 2.93 Ghz Intel six-core processors and 96GB of memory – including dual port 10G FCoE can be had for about $11,700 from the HP website not including any discounts. I did find some pricing for Cisco on line and this Blade could sell for $13,665 – discounted from an MSRP of $26,407.
But looking at this case study, out of the forecasted $187,000 in capital cost savings, as much as $150,000 comes just from the reduction in Microsoft Windows Server licenses and other software license savings…..leaving Cisco responsible for about $30,000 of the savings – so maybe 15% of the total. Is that the best they can do?
logo





So if it really does save 15% above the licensing costs and (as it claims) it saves the customer a significant amount ($1678) per logical server going forward, cuts per server management costs from $1575 to $80, and offers them the flexibility to deploy B250 blades capable of 384 GB of memory with just 8GB DIMMs, what is it that you feel is another ‘snow job’?
James,
Are you saying that Cisco unique product capability is responsible for the significant (75%) cost savings?
Out of the claimed 75% savings in this Cisco case study…..at most , Cisco could claim 15% savings….and that isn’t clear. To suggest that Cisco had a meaningful unique contribution to the 75% savings IS the “Snow Job”. I expect the savings could be even higher with HP or DELL blades….it seems that virtualization was the basis of this savings…not the hardware platform.