I would normally ignore yet another “overview” of clouds, but being McKinsey it will get read by executives and several of their generalizations about “not being cost effective for large enterprises” are just plain misleading.
McKinsey calculates TCA (Total Cost of Assets) for a data center as $ 45 per CPU a month. They use that as a basis to compare against amazon’s compute benchmarks and conclude it if you were to outsource your entire compute demand to them it would be pricier than an internal DC.
Benchmarks I have seen from Gartner for managed services ( their version of McKinsey’s TCA) for even a simple DC environment range from $ 200 to 1,000 per server per month. They are even higher for more complex environments. Compared to that amazon should be competitive for most scenarios.
Besides, most CIOs I know are looking at clouds as an incremental, not a replacement, DC strategy. There are compelling capex and provisioning time reasons to use clouds for new, not necessarily, replacement sourcing.
McKinsey also does not compare storage costs – amazon versus current DC costs. If you needed a petabyte of storage, amazon would sell it at $ 1.44 million a year. Add fees for transfers in and out and a reasonable estimate would be $ 2 million a petabyte a year. I challenge McKinsey to amortize EMC, IBM or other storage common in most DCs (the hardware, software, maintenance, extended service) and get anywhere close to that annual number.
McKinsey then says “ enterprises set their SLA uptimes at 99.99 or better, which cloud providers have not yet been prepared to match”.
“Who started that myth? Other than some sensitive and global apps, there is plenty of downtime in corporate data centers and those of big outsourcers like IBM and EDS.
In a given year (unlike a leap one like this year), we have 525,600 minutes. To meet a 99.99% uptime, the system could only be down 52 minutes - less than an hour - in the year. I can tell you most corporate data centers have scheduled downtimes which exceed that every month, if not every week.”
McKinsey recommends “Rather than create unrealizable expectations for “internal clouds,” CIOs should focus now on the immediate benefits of virtualizing server storage,network operations, and other critical building blocks”
CIO should be doing that – but they should also use cloud benchmarks to dramatically renegotiate downwards much of their current infrastructure and application costs. And increasingly move to vendors which been delivering cloud services for years now – like the ones being showcased in my “Cloud Pioneer” series.
In other words they should do what they typically do when a senior executive hands them a McKinsey report on most IT matters. Politely acknowledge it and and then ignore it.
Update: I neglected to mention, and was jogged by Nara's post, that the McKinsey study talks about salesforce, Zoho and other SaaS as "cloud services" but makes no attempt to compare them against costs of on-premise SAP, Oracle, Microsoft applications. I am sure you could go back a few years in many of McKinsey's clients and see they helped justify those very software investments :)
I would normally ignore yet another “overview” of clouds, but being McKinsey it will get read by executives and several of their generalizations about “not being cost effective for large enterprises” are just plain misleading.
McKinsey calculates TCA (Total Cost of Assets) for a data center as $ 45 per CPU a month. They use that as a basis to compare against amazon’s compute benchmarks and conclude it if you were to outsource your entire compute demand to them it would be pricier than an internal DC.
Benchmarks I have seen from Gartner for managed services ( their version of McKinsey’s TCA) for even a simple DC environment range from $ 200 to 1,000 per server per month. They are even higher for more complex environments. Compared to that amazon should be competitive for most scenarios.
Besides, most CIOs I know are looking at clouds as an incremental, not a replacement, DC strategy. There are compelling capex and provisioning time reasons to use clouds for new, not necessarily, replacement sourcing.
McKinsey also does not compare storage costs – amazon versus current DC costs. If you needed a petabyte of storage, amazon would sell it at $ 1.44 million a year. Add fees for transfers in and out and a reasonable estimate would be $ 2 million a petabyte a year. I challenge McKinsey to amortize EMC, IBM or other storage common in most DCs (the hardware, software, maintenance, extended service) and get anywhere close to that annual number.
McKinsey then says “ enterprises set their SLA uptimes at 99.99 or better, which cloud providers have not yet been prepared to match”.
“Who started that myth? Other than some sensitive and global apps, there is plenty of downtime in corporate data centers and those of big outsourcers like IBM and EDS.
In a given year (unlike a leap one like this year), we have 525,600 minutes. To meet a 99.99% uptime, the system could only be down 52 minutes - less than an hour - in the year. I can tell you most corporate data centers have scheduled downtimes which exceed that every month, if not every week.”
McKinsey recommends “Rather than create unrealizable expectations for “internal clouds,” CIOs should focus now on the immediate benefits of virtualizing server storage,network operations, and other critical building blocks”
CIO should be doing that – but they should also use cloud benchmarks to dramatically renegotiate downwards much of their current infrastructure and application costs. And increasingly move to vendors which been delivering cloud services for years now – like the ones being showcased in my “Cloud Pioneer” series.
In other words they should do what they typically do when a senior executive hands them a McKinsey report on most IT matters. Politely acknowledge it and and then ignore it.
Update: I neglected to mention, and was jogged by Nara's post, that the McKinsey study talks about salesforce, Zoho and other SaaS as "cloud services" but makes no attempt to compare them against costs of on-premise SAP, Oracle, Microsoft applications. I am sure you could go back a few years in many of McKinsey's clients and see they helped justify those very software investments :)
McKinsey's Dark Clouds
A reader pointed me to McKinsey’s report on cloud computing
I would normally ignore yet another “overview” of clouds, but being McKinsey it will get read by executives and several of their generalizations about “not being cost effective for large enterprises” are just plain misleading.
McKinsey calculates TCA (Total Cost of Assets) for a data center as $ 45 per CPU a month. They use that as a basis to compare against amazon’s compute benchmarks and conclude it if you were to outsource your entire compute demand to them it would be pricier than an internal DC.
Benchmarks I have seen from Gartner for managed services ( their version of McKinsey’s TCA) for even a simple DC environment range from $ 200 to 1,000 per server per month. They are even higher for more complex environments. Compared to that amazon should be competitive for most scenarios.
Besides, most CIOs I know are looking at clouds as an incremental, not a replacement, DC strategy. There are compelling capex and provisioning time reasons to use clouds for new, not necessarily, replacement sourcing.
McKinsey also does not compare storage costs – amazon versus current DC costs. If you needed a petabyte of storage, amazon would sell it at $ 1.44 million a year. Add fees for transfers in and out and a reasonable estimate would be $ 2 million a petabyte a year. I challenge McKinsey to amortize EMC, IBM or other storage common in most DCs (the hardware, software, maintenance, extended service) and get anywhere close to that annual number.
McKinsey then says “ enterprises set their SLA uptimes at 99.99 or better, which cloud providers have not yet been prepared to match”.
Really? As I wrote last year
“Who started that myth? Other than some sensitive and global apps, there is plenty of downtime in corporate data centers and those of big outsourcers like IBM and EDS.
In a given year (unlike a leap one like this year), we have 525,600 minutes. To meet a 99.99% uptime, the system could only be down 52 minutes - less than an hour - in the year. I can tell you most corporate data centers have scheduled downtimes which exceed that every month, if not every week.”
McKinsey recommends “Rather than create unrealizable expectations for “internal clouds,” CIOs should focus now on the immediate benefits of virtualizing server storage,network operations, and other critical building blocks”
CIO should be doing that – but they should also use cloud benchmarks to dramatically renegotiate downwards much of their current infrastructure and application costs. And increasingly move to vendors which been delivering cloud services for years now – like the ones being showcased in my “Cloud Pioneer” series.
In other words they should do what they typically do when a senior executive hands them a McKinsey report on most IT matters. Politely acknowledge it and and then ignore it.
Update: I neglected to mention, and was jogged by Nara's post, that the McKinsey study talks about salesforce, Zoho and other SaaS as "cloud services" but makes no attempt to compare them against costs of on-premise SAP, Oracle, Microsoft applications. I am sure you could go back a few years in many of McKinsey's clients and see they helped justify those very software investments :)
April 16, 2009 in Industry Commentary | Permalink