November 20, 2019:
The US Department of Defense recently (25 October) selected the winner of long overdue JEDI (Joint Enterprise Defense Infrastructure) contract. Microsoft will get $10 billion over a 10–year period to provide a single cloud computing network for all online Department of Defense operations. This announcement was a quite surprise because many favored Amazon Web Services instead, which earlier was a contractor for some smaller Department of Defense online storage “clouds”. Whatever the case Microsoft has the resources to do the job and a major incentive (attracting more customers) to do it right. The major problem with the Department of Defense is not the large number of cloud users but the large number of subagencies that possess considerable autonomy and political support. Dealing with this toxic political environment has been the cause of many major Department of Defense projects failing, or succeeding but at much greater time and cost. Then there is the problem of supporting a lot of users who will be in combat or in a combat zone. The Department of Defense is not your typical large cloud customer.
The modern battlefield is becoming a more and more complex and deadly environment when information and its processing speed is as important if not even more than modern equipment. That would include tanks or combat aircraft. As a result, many countries have come to appreciate the need to upgrade their doctrine accordingly including China, Russia and France.
The U.S. armed forces branches have been trying to improve cooperation by themselves. Past examples include Air-Sea battle concept from 2011 or the U.S. Army 2016 Multi-Domain Battle. These more or less successful attempts have created many systems and databases which are not compatible with each other. This mess forced U.S. troops to use one system to get information about adversary, another system for weapon selection and a third for situational awareness.
Worse, American armed forces have many data centers together with the command, control, and communications systems. While some are quite new, others still use outdated tape or even storage or even 8-inch floppy disks. Earlier in 2019 the floppy disks were finally put to rest. In theory, data are there in both cases but access to them was, to say at least, cumbersome in addition to slow reaction time.
The JEDI contract will be the backbone for the new system which aims to bring all data and a few hundred smaller clouds into one system with clear standards. For example, the Pentagon has been testing Project Maven. This system is described as using Artificial Intelligence but in reality, it is a machine learning algorithm. Maven is designed to perform the mundane but essential process of looking at the massive amounts of video from reconnaissance aircraft and other sources and make sense of it all in a predictable and rapid fashion. The lack of an enterprise cloud solution had slowed the Project Maven learning process. The project staff simply weren’t able to get enough data for analysis fast enough because archived data first needed to be processed into a common format. This additional work wouldn’t have been needed if there been an enterprise cloud solution.
The other benefit of the central JEDI cloud will be the ability to move data from any part of the force from anywhere in the world. All five recognized domains: land, sea, air, space and cyberspace will be connected and gain the ability to share information in an accessible form. Besides the central JEDI cloud, there are plans for smaller or even portable data centers with AI capabilities. For example, when the ground forces need air support they could call it much faster thanks to a cloud/AI combination. In this scenario, the AI would check the closest available and suitable weapon system (fighter or artillery) and guide it to needed area. This could be done in mere seconds.
The JEDI is the first and the most important step to build U.S. networked armed forces. The next big challenge will be the migration of existing clouds to it and developing mature enough artificial intelligence or machine learning algorithms needed to sort and synchronize data with new communications systems or new equipment. It should be noted that the recently (2018) created the Joint Artificial Intelligence Center will have much to do in the coming years supporting the establishment of JEDI. ---Przrmyslaw Juraszek