Sarbjeet Johal · @sarbjeetjohal
12th May 2021 from TwitLonger
#Cloud musings... (in response to @mthiele10's Linkedin post)...
We will be hybrid for a "long" time to come for sure, but net-new spend will keep shifting to "value based" pricing in infrastructure (regardless location) as well as applications (SaaS) space. That trend is irreversible. Reminder, #publiccloud has invaded your DC, yep, and they are just getting started! Also, remember that on-prem is still 80+% of spend. As Tim Crawford said, so much opportunity...when distributed appropriately. I believe those #SpendAppropriations are the hardest decisions for IT leaders.
My commentary on Mark Thiele's thesis: this idea (of on prem infrastructure over cloud) holds true if enterprise workloads were homogenous (like Google's workload). Biggest problem is in enterprises is #UtilizationRates on storage and compute are very low, there are technological as well as organizational structural issues at play (silos) for this low utilization rate. On technical side, infrastructure isn't as #divisible as it should be. My team's DC audit work spanning more than million servers (and attached mumbo-jumbo) proved this that if by going to cloud we can bring up utilization rate up by ~20% (which is easy IMO) and reduce (new or marginal) procurement time by 90% it's better than saving 30% on gear you are buying (to own). When supply curve hugs demand curve, you aren't wasting much. Yes, cloud premium is there but it pays for itself... big time. BTW if you don't know what you are doing, any tooling or business models can hurt you, that includes public cloud... as Grady Booch says, "a fool with a tool is still a fool".
Having said that, pendulum may swing back in long term, but not in next 3-5 years. [context: I am talking about where net-new spend is going... owning vs paying per use, regardless location].
Long live #PublicCloud! 😉