Utah Data Center: Inside the NSA's Hard Drive

[00:00:00] Speaker 1: this video is sponsored by keeps more on them in a bit in may of 2014 an it contractor working for the national security agency or nsa fled the united states after leaking terabytes of secret top secret and other classified documents in this leak programs such as prism which conducted intrusive surveillance on social media and search engine servers and dishfire which captured and stored text messages and related data and metadata were revealed to the public included within the many terabytes worth of leaked documents was key information about the utah data center the nsa's main data storage and backup facility since then the nsa has attempted to get ahead of the leak and regain control of the narrative with greater or lesser success and one of their tactics has been to compile and publish this leaked information along with explanatory notes whether or not this has worked for the nsa what it does mean is that we're now in possession of some of the only hard information available about one of the most secret places on the planet the nsa is what's known in the intelligence community as a sigint or signals intelligence agency what this basically means is it's their job to intercept and collect communications and other signals and electronic and radio signature information before either storing it for future use or processing the data into usable intelligence in the days before cell phones and the internet signals intelligence was mostly concerned with radio satellites telephony and the various electronic and radio emissions of military equipment the rapid explosion of information technologies from the late 20th century onward erratically transformed the shape and function of the nsa catapulted onto the front lines of the cyber battlefield the nsa very suddenly became the custodian of massive amounts of data as advances in computing and the surveillance capabilities associated with this brought them more and more material all of which had to be stored while the government tried to figure out what to do with all of this intelligence the rising threat of terrorism in the post 9 11 world accelerated this process and in the years leading up to snowden's and other whistleblowers revelations the nsa always faced with gargantuan levels of data storage requirements this led to a secretive program first revealed by journalists combing budget reports to build a massive data center in utah according to snowden this project was initially called the massive data repository which was later changed to mission data repository according to the official nsa site however it's still called the massive data repository as for the choice of utah as a location the state has long been a hub for digital technology facilities the university of utah housed the only arpanet node outside of california and some of the country's most prestigious computer science programs were taught at that institution arpanet was a precursor of the internet created by the advanced research projects agency a department of defense research agency which was eventually to become darpa utah is also home to an area nicknamed silicon slopes playing on silicon valley which houses facilities for companies like microsoft oracle and adobe additionally quite a few highly secretive programs have been based in utah attracted by the wealth of remote isolated and easily defensible locations available in the state the state government itself pitches hard for these projects as well pointing out the general lack of natural disasters and their willingness to cut sweetheart deals for utility costs and infrastructure builds and this is why utah is home to facilities for training drone pilots and testing chemical and biological weapons among others so it was that in the mid-2000s the program chose a remote patch of flatland with a nice mountain view of the edge of camp williams a national guard based just south of bluffdale for the construction of its massive data repository have you ever wished yet someone to help you with your hair loss look i certainly have two out of every three guys are going to experience some form of hair loss by the time they're 35 but keeps offers clinically proven and research-backed treatments that not only stop hair loss but can also improve hair growth and give you healthier fuller looking hair plus their treatments are recommended by licensed medical providers and come with the convenience of being delivered directly to your door and it's all at about half the cost of a traditional pharmacy not only do they deliver right to your front door but keeps provides you with a network of expert medical advisors prescribers and care specialists who can help you make your hair goals a reality each treatment plan comes with unlimited messaging support with keeps licensed medical providers so you can connect at any time and ask all your hair related questions look whether you're looking to prevent or slow your hair loss stimulate hair growth for a thicker and fuller head of hair or simply want to maintain and take better care of the hair that you have keeps has got you covered when it comes to your hair needs from their clinically proven treatments to an award-winning all-natural thickening shampoo and conditioner system they have everything you need to reach your hair goals so don't wait any longer take control of your hair journey with keeps today hair loss stops with keeps to get 50 off your first order go to keeps.com mega projects or click the link in the description that's keeps.com mega projects for 50 off their clinically proven hair loss treatments and now back to today's video in 2010 the u.s army corps of engineers who were to supervise construction awarded the build contract to a consortium of firms headlines by balfour betty dpr and the amusingly named big d belong along with sorry childish and the amusingly named big d along with the architectural firm king stubbins construction began soon after the 1.2 billion dollar contract was awarded over the period of the build the massive tract of land was cleared fenced and excavated with a small army of security contractors and specialist staff called in to help initially security for the site proved problematic with guards and agents grumbling that the specially hardened chain and barbed wire fence was too low and too see-through they were also unimpressed with the coordination of access control as the new roads were built and opened ahead of adaptations to the patrol and guard patterns journalists would wander into the site either accidentally or deliberately failing to spot the signs warning them that trespassing was prohibited and they were on official federal property the security staff complained the signs were too small and cashmere hill a journalist for forbes reports blowing straight past without noticing them before she was stopped and detained for more than an hour for taking pictures at one point press reporting on the site inspired two sovereign citizens to roll up to the gate and demand an explanation for the construction interestingly during the time mrs hill was detained she pointed out there wasn't really any point making her delete the pictures she took as the whole site was basically open to view from the ground or the air the special agent who had detained her cryptically responded that only less than half of the site was in view if this throwaway comment is to be taken seriously it means that a significant portion of the utah data center is underground this is interesting as none of the official or unofficial documentation mentions an extensive underground complex as part of the deal cuts with utah for sitting the data center in bluffdale new items of infrastructure were built for the site developments which were welcomed by the government chief among these was a four mile water pipe and a three million gallon water tank as the open date of 2013 approached and as an increasing amount of more sensitive installation was taking place security was tightened significantly and most of the random incursions stopped in that same year however the series of unexplained and highly destructive power surges reached havoc with the nearly finished data center and the opening had to be delayed for a year it's been estimated that each surge of which there were 10 reported caused a hundred thousand dollars worth of damage the official explanation for these power surges was that the electrical system designed by king stubbins and called upon to provide unprecedented requirements had interacted with itself in unpredictable ways whatever the truth of the matter the electrical problems were solved the data center the official name for the utah data center is the intelligence community comprehensive national cyber security initiative data center as this is very catchy the site is commonly known by its code name bumblehive the official function of the bumblehive is to provide a centralized repository of immediately accessible data to authorized intelligence customers what this means is that other intelligence and law enforcement agencies including those of allied countries can apply to the nsa for access to the data they've collected this can be helpful in terms of fusing intelligence a process whereby signals human and other types of intelligence are synthesized in order to provide a more complete picture of observed activity according to the nsa the heaviest staffing burden of the utah data center revolves around controlling access to this data highly cleared officials are tasked with processing data requests and determining whether or not the requester has the appropriate clearance and legal authority to see what they're asking for the bumblehive takes up an area of 1.3 million square feet or more than 12 hectares with a hundred thousand square feet being the actual data center itself the rest of the area is taken up with support administration and ancillary facilities in terms of the power the center soaks up 65 megawatts a huge amount of electricity costing about 40 million dollars per year there are 20 buildings on the site these range from central administration buildings to firefighting facilities as well as a dedicated pump house emergency diesel generators an electrical substation water treatment plant refrigeration plants and security related structures such as vehicle inspection bays and visitor control facilities the 60 emergency generators are said by the nsa to be capable of providing 100 power backup for three days other sources report that they can power the core data center continuously for an entire year there are quite a few conflicting media reports about the data capacity of the bumblehive with estimates topping out at several zettabytes so zettabyte by the way is basically a hundred million terabytes the nsa is tight-lipped about the total capacity saying that the actual number is classified for national security reasons they do however quote utah governor gary r herbert when he boasted on his blog that the bumblehive would have a yachtabyte's worth of capacity a otabyte is a thousand zettabytes or one trillion terabytes context the entirety of the data created consumed and stored by the whole world in 2020 was 64.2 zettabytes so does this mean that governor gary accidentally let some information slip and the nsa is tacitly acknowledging it or is it clever misdirection either way all the nsa will officially say is that the data center was designed with future capacity expansion in mind and that the current and future capacity can be measured in a lot of bytes very clever nsa there's also some arguments as to whether or not the utah data center is the biggest data repository in the world there is no argument about the fact that china has the physically largest data center china's people's republic telecom data center is widely known as being the biggest in the world but it is unknown how much data it holds having said that the law in china mandates that all of the data associated with its 1.4 billion people must be held on government servers there's also a complete lack of meaningful privacy protections in the chinese state which means that theoretically there is no limit to what information can be collected by the government this combined with the fact that china is set to produce more data than the us by 2025 makes it highly likely that the bumble hive is at best in second place which isn't necessarily a bad thing according to some sources the data centers drives a solid state meaning that they require less power and cooling than spinning tape drives as they have fewer mechanical moving parts it's thought that the data being stored here is being mirrored so it can be instantly recovered in case of failure the data is also striped which is the practice of storing consecutively sequence data in different locations so that failure in one location won't wipe out the whole string it's important to note that these details informed speculation as the inner workings of the bubble hive are necessarily classified the whole center is run by a massively parallel cray xc30 supercomputer massively parallel processor arrays are a collection of central processing units or cpus in which hundreds of thousands of cpus and ram banks are integrated into a single circuit if we think of each cpu as a brain providing a certain amount of the way a massively parallel processor array works is by pooling hundreds of thousands of brains together so that they can pass work to each other over a specifically built interconnect this allows large and complex problems to be segmented with each segment being figured out simultaneously or in parallel in order to finish the whole problem more quickly the massively parallel cray in the bumble hive can perform a hundred thousand trillion calculations every single second the nsa has stated that it's currently working on cracking something called the aes algorithm aes stands for advanced encryption standard and is the encryption algorithm used all over the world 256-bit aes encryption is highly secure and it would take the bumblehives cray a few hundred years to brute force a single instance of aes-256 encryption the nsa's computer research facility at oak ridge has been working on this problem as new supercomputing centers revealed in budget papers indicate that they may well be on the way to building an exoflop computer an exoflop machine would be capable of breaking aes encryption within reasonable time frames until the day that they have one the nsa says that one key function of the bumble hive is to store captured encrypted data against the possibility of cracking it in the future amusingly the utah data center has a two-star rating on google reviews a great many of these accuse the nsa of stealing utah's water the western usa is currently in the grip of mega drought and water supplies are critical all up and down the midwest especially in those places which depend on the colorado river the reasons for this are complex and stretch back a long way utah consistently has some of the highest water usage figures per capita in the country and over the years states along the colorado have played fast and loose with water supply subsidizing water intensive agriculture and making environmentally questionable decisions this is especially critical in utah which is one of the driest states in america and the great salt lake after which salt lake city is named has been visibly shrinking for several years this has brought the water shortages into critical relief and the nsa having a fair amount of negative profile over the years has become something of a target on this issue this is because data centers use a lot of water in order to run their computers or supercomputers a huge amount of cooling needs to be done and the most effective way to do this is through evaporative water cooling according to bluffdale city records the utah data center used 128 million gallons of water between june 2021 and may 2022 this is the equivalent to the consumption of 2 800 individuals the facebook data center which is only slightly smaller used 13 million gallons for the same period this drain on the city's water is a contentious point for those who live in utah many of them believe that the state and city governments have made a poor deal exchanging short-term gain for long-term drainage of vital resources the nsa for its part says that it's aware of the issue with water consumption and that it's working hard on innovative dry cooling techniques to reduce their impact it says that it's already reduced its expected water consumption by a third at returning 34.5 million gallons of water to the jordan river they point out that their used water which is purged from the system when it's no longer usable for cooling is used for irrigation in bluffdale they also politely point to the 3 million gallon tank and pipeline that they paid for this seems to have done little to mollify many residents of utah some of whom are passionately opposed not just to the bumblehive but to all the data centers current or planned at a time of water shortage and economic turmoil the plant 300 acre google campus which will also be cited in utah is a major point of contention and the nsa's facility is frequently brought up in the discourse around this and other planned developments the other great controversy has to do with the issue of what the nsa actually does it's hard to believe but there was a time not so long ago where very few people had even heard of them the nsa was an obscure highly technical organization which hardly anyone not even hollywood scriptwriters really thought of as sufficiently interesting to mention this all changed after the snowden and other leaks however in the nsa has emerged as the bogeymen of the surveillance state some of this reputation is absolutely justified and is clear from the leaks and investigative journalists that the nsa surveillance program had massively overreached itself in the wake of 9 11. while claims that the agency is watching and recording every digital communication made in the united states are highly questionable it is understandable that so many people believe this to be true when the utah data center was in its final year of construction two nsa whistleblowers william binnie and kirk weiber took to the press with claims that the nsa was capturing not just the metadata of all domestic communications but also the content metadata for those who don't know is data concerning things like time place and participants in a conversation but does not include the conversation itself binnie who was a technical director at the agency argued that the massive capacity of the bumblehive could only be for capturing all of the content as metadata storage is much lighter and would only need a warehouse size facility against this is the official nsa position they have repeatedly asserted that the nsa is constrained by the law and that it has oversight from every branch of government a fact they say makes it impossible for the agency to overreach its authority they also spent some time blackening the reputations of binnie and weiber saying that they'd been losers in internal politics were in bitter malcontents and had been outside the agency for so long they couldn't possibly know what was going on despite these denials or perhaps even because of them a great many authoritative sources report as a matter of course that the utah data center is the repository for the nsa's collection of every u.s citizens text messages emails phone calls and payment and location histories the nsa says it's merely a repository for data collection which has been strictly targeted as well as suspicious encrypted communications and that access to this data is strictly controlled however we might choose to believe on this one it's still a fact that the utah data center or the bumble hive is one of the most impressive achievements of the information age for good or for real