Re: Whole single vendor thing was a scam from the beginning
This was presented as a single vendor solution by Oracle because it suited the picture Oracle were trying to paint. It's not.
a) this is for part of the DoD annual infrastructure spend (approx $1bn/year out of the ~$19bn-$25bn annual projected DoD budgets)
b) other vendors will definitely be involved, this is for the preferred supplier for GovCloud services and will likely reduce the number of data centres to AWS/Azure GovCloud facilities PLUS one data centre per state PLUS special project data centres - a total of ~220 data centres will be reduced to <90 with the majority of big providers still running some facilities.
c) the DoD has separately awarded contracts to Microsoft for Office365-type functionality. Awarding Microsoft the JEDI contract brings them closer to the single vendor model you mention.
d) the DoD currently manages multiple suppliers AND they have been screwing the DoD. Estimated "cloud" spending for 2019 is $1.2bn (note the saving if that was to reduced to $1bn...) and an estimated $1.5bn in 2020 with the expected cost of DoD GovCloud services without the JEDI contract in the order of $17bn-$22bn vs $10bn.
e) Microsoft can't deliver the required services to the DoD on day 1 - AWS can. This means some of the proposed savings will only happen if the contract goes to AWS because Microsoft are still building out their GovCloud facilities.
f) AWS GovCloud systems will need to be moved to Azure versus benefitting from lower negotiated rates. As AWS was the de facto GovCloud standard for the DoD for the last 2-3 years because of a lack of alternatives, this will reduce the savings and provide additional migration work to get systems into Azure.
g) This contract is primarily around cost control - most of the vendors have been screwing the DoD for years and this is the DoD screwing them back. Hence the legal fights.
h) for in-house vs outsourced, the DoD outsourced the majority of their data centre facilities in the mid-to-late 90's. That ship sailed a long time ago.
i) for security, the work Google and AWS have done to address existing security issues with DoD infrastructure has significantly increased security in the last ~10 years. The existing vendors said their disjoint security models provided adequate protection but upon review it turned out to be a very optimistic view. Google and AWS provided new models and comprehensive reviews that eventually fed into GovCloud as standards - the costs of bringing existing vendors up to these standards has been one of the primary drivers for reducing the number of vendors who have DoD access - they were overcharging and under delivering.
TL;DR: AWS were lowest cost AND best vendor for JEDI based on the requirements. Microsoft were also closely aligned with the DoD, just not for JEDI. Instead of saving money, JEDI under Azure will result in less vendor diversity and significantly reduced cost savings