00:00for example YouTube would do it was it would have not only your data in one
00:03location but it would have the entirety of your data in multiple locations
00:06around the world have it cached right and so they use way more data than they
00:11should for per video but they do it because it's a service that really needs
00:15like streaming access immediately so that's the whole idea why we're
00:19encrypting the file with your own private key in your device so that means
00:25the when the file is transferred to us it's already encrypted with your private
00:28key and you need your private key to open that file so for storage all of them
00:34majority of them follow the S3 standard which is the Amazon S3 standard that's
00:38what it is our system is compliant to S3 standard so you just need to change those
00:44pointers and you can really migrate your data easily
00:58Tangelic Talks your go-to podcast from Tangelic where we dive into the vibrant
01:03world of clean energy development sustainability and climate change in
01:07Africa we bring you inspiring stories insightful discussions and
01:11groundbreaking innovations from the continent making waves in the global
01:15community tune in and join the conversation toward a brighter greener
01:19future let's get started
01:23welcome to Tangelic Talks a podcast at the intersection of energy equity and
01:28empowerment with your co-hosts Victoria Cornelio and Andres Temes in today's
01:32episode we are diving into one of the most invisible but rapidly growing climate
01:37challenges of the digital age data every search every AI model every cloud upload
01:43lives somewhere physically and increasingly those massive centralized data
01:46sensors are consuming enormous amounts of energy and resources so our guest today
01:51Murphy John from StoreX Network a platform exploring how decentralized
01:56storage could reshape the future of digital infrastructure is going to help us
02:01understand a bit more about digital autonomy resilience and the
02:04environmental cost of the internet itself welcome Murphy hey thanks for having me
02:09really excited to be part of the podcast no this is great honestly we keep
02:14talking about this this season and I think we really needed to nail down this
02:18conversation because the internet feels invisible many times you know every drive
02:24feels like it's in the ether and we don't really take into account how that's
02:29happening where is it you know so most people don't think about the
02:37environmental footprint of data can you give us a bit of an idea of how big of an
02:41issue these centralized data sensors are yes actually I'll start with background so
02:46first let me talk about the data that we are generating and then we get on to the
02:49storage of it so if you see the kind of amount of data that we've been generating
02:54over the period of the years like it has been enormous so just some statistics the
03:00amount of data that we generated over the past five years is far more than the
03:04data that we have generated in the past 20 or 30 years put together that means we are
03:09generating a huge amount of humongous amount of data and this data needs to be
03:13sourced somewhere or the other now what what's what's really driving this data
03:17because there are a lot of IOT devices smart devices all over the world
03:21driving smart cars coming in and a lot of autonomous software and of course AI is one
03:26of the very big chain that plays a very key role in this so what happens is as
03:31soon as you start generating so much amount of data you need a store in your
03:35location to store it and you need to store it very securely and the reason why I'm
03:41talking about data is we we really have this feeling that we are living in the
03:45data economy right now so data has become really very crucial for every
03:49organization and so much of data being generated so much of the output is
03:54generated I think it's a very critical and a very crucial factor in the life so
03:58what happens is this data actually everybody has if you talk to a non tech
04:03person he'll see the data is on the cloud he doesn't know where it is so this time
04:07yeah it's in the ether yes yes so it's it ends up on one of the physical machines
04:13which is located somewhere at some location and that's that's what is the
04:18driving challenge that we have today because this all data is actually
04:21residing in some data centers and this end these data centers are one of the
04:26highest consumers of electricity and all the other resources put together and if you
04:31see just the data center the data center is almost like conceiving not only
04:35conceiving huge amount of electricity water and all those resources around but
04:41they they also need they are also underperforming so that means if you do if
04:46you do an access or we do an audit of all the data centers they are all being
04:51underutilized so if there is a huge data center that has been set up it's
04:55actually underutilized so that's where the huge issue comes up because they
05:00cannot scale up to the full level and there's a huge amount of electricity and
05:04all this resources that's being wasted so most of them but you see there is there
05:09is a demand which is growing and there's the this data centers which are lying ideal
05:14so when we talk about data centers we always talk about this hyper scalers like
05:19Amazon, Google, Microsoft all of these guys are spread but there are multiple data
05:24centers multiple service providers spread worldwide so what happens is we are in
05:29the mission so we we found that this data is something that is very critical and
05:34this data center is consuming so much of electricity wasting so much of
05:37electricity because of the under utilization so we kind of tend to bring
05:41these things together so what we do is we utilize storage which is located in
05:46this and such underutilized IDCs and then we store the data so we are not
05:51setting any green field data centers we are not setting up any additional data
05:55centers any resources what we do is we have a community member so we have a
06:00community of storage node operators who are spread worldwide so we have like in
06:03multiple geography in multiple locations they go and source this data centers which
06:08are underutilized and they bring in the resources to us where we store this data so
06:12instead of storing your data in a single location we spread your data and spread it
06:17across to them and because we're doing in this kind of what helps actually the
06:21the other additional advantage of using us as it is because we're not storing
06:26everything together in a single location which was traditionally how the data is
06:30been stored over so many years if you go to any hyperfilis they just take your
06:34file and store it in one single location but what we do is we break your file
06:38first we encrypt your file we break it into multiple pieces and then we just
06:41store this encrypted fragments across multiple locations so what happens is I'm
06:46spreading out my data I am like decreasing the risk which is associated
06:50with single file of storage and then as well as utilizing storage which is
06:55located on or the data center which is underutilized so that's that's how the
06:58economy goes yeah that's really really cool so you're maximizing the usage of
07:03data centers that already exist so that so that we're not wasting resources what how
07:10when you when you say that of is the are the files themselves being split and
07:16encrypted so that you basically don't have that's super cool so it's very it's gonna be
07:23very hard for someone to get access to that data and to use it maliciously and if they get
07:31access to one data center then they don't have access to the data anyway right they have to get
07:36access to the whole thing and then decrypt it that's super interesting now can you can you tell
07:42us a little bit about a little bit more about before I can start asking questions about your
07:46product like is this an enterprise product is this consumer product like what type of data do you guys
07:52store mostly so we have board we have offerings which is there for B2C customers so the end users
07:58can just come on log into the website register we have a 2gp free for life plan so they can
08:04just register
08:04sign up and use started use they don't have a part of credit cards or anything like that so you
08:09can
08:09just sign up and start using it so that's the B2C configuration and we have a B2B setup where we
08:14work with enterprises around the world so who is looking at storage use cases they have multiple
08:20use cases requirement we work with them we design systems around them and then we give this system
08:26design for them so we work operated in both scales B2BS and this B2C so one of one of the
08:31issues that I
08:32can see with that is latency right to access your data if it's being split up among multiple locations
08:40in fact one of the ways that like like you say the hyperscalers deal with this now you guys don't
08:48have
08:48to do this but the way for example YouTube would do it was it would have not only your data
08:53in one
08:53location but it would have the entirety of your data in multiple locations around the world have
08:57it cached right and so they use way more data than they should for per video but they do it
09:03because
09:04it's a service that really needs like streaming access immediately so do you guys is it like it's
09:11okay to sacrifice just a little latency because it's not a streaming platform right you just need
09:15access to your data or do you guys have ways to mitigate that yeah so so so the other guys
09:23the
09:23hyperscale is what they do is YouTube is a good examples they have pop locations which are at
09:28multiple location multiple geography and then they have hearings with the ISP that ensures that this
09:33kind of leaders is not there because specifically they're working with videos and streaming is like
09:38the main so people don't want to wait when they're looking for some streams for us it's like business
09:43data so what the the advantage that you're working on or the best advantage that we have is because
09:50we're spread out across multiple geographies we can offer them very high level of reliability and
09:55resiliency for the data so it's not like just one location one thing and and the algorithm is
10:02designed in such a way that it will always look at the nodes which are really available so what one
10:07thing
10:08that we do is I will just come back to the way how we do the entire engineering around the
10:13product
10:13so whenever you're using storex to secure your data so what we do is in the user's laptop or device
10:20itself we will encrypt the data with their own private key so that means it's encrypted it uses
10:25device only and then the encrypted file goes to the server where it's broken down into multiple
10:30fragments then we replicate those fragments and then we distribute it so actually what we do is
10:35your file when we are dividing it into fragments we create multiple copies and these multiple copies
10:41are located different geographies so suppose there is an A file you have like one two three four five
10:47one file would again be divided like replicated three times or four times depending upon what is the
10:53nature of the file and this will be residing in three different geographic locations so whenever
10:58you're requesting a file the system will know whether file is and then it will know which is the nearest
11:05location from where you can get this file so I won't go to the remote this location to get this
11:09file I
11:10my system will know that this is located suppose you are in Berlin and you have a look you have
11:15a
11:15location in Frankfurt that means that the file you can fetch the file very fast go to the Frankfurt
11:20location get this French those so that's how it operates in the engineering okay so you guys do do
11:26have it set up so that there are multiple copies of it but you guys mitigate the environmental impact by
11:32like you say using data centers that already exist that are under that are underutilized right yes
11:39yes okay is there is you know there's a difference between data centers right like there's storage data
11:46centers and then there are compute data centers yes um how what is the environment the difference in the
11:52environmental impact right between something like a storage data center and the compute one uh like in how
11:59much cooling they need how much power they consume because I think that's important to know right like
12:04not all data centers are the same yeah so uh so because uh the compute is more intensive it requires
12:11more
12:12energy and more efficiency to operate of course because of the new uh the new kind of uh development that
12:18is coming in because of and all those kind of development you see this consumption coming up but if you
12:24compare
12:24storage with computing I think the the computer part would be much more energy intensive and it will
12:30be utilizing more resources compared to the storage because storage is just dumb storage you don't do
12:34processing around it so and in computers you're actually processing a lot of data yeah that that makes
12:39sense right they still they still need to there's still a lot of heat that needs to be managed and
12:43stuff like that right so you need a lot of cooling and and such so that would be a lot
12:48of the cooling capacity
12:49I think is the big one for for data storage yes so that that that was that's really a big
12:55amount of
12:56HAVC that's required cooling where do you like see the industry going do you see it going in the
13:03direction that you know data data is really interesting because there are multiple facets of
13:10the morality to it right you have the server itself which generates heat it consumes a lot of energy
13:15and then you have the aspect of the data itself for example you guys I imagine you said you said
13:21you
13:21deal a lot with enterprise b2b clients and stuff like that right and also consumers um but is your
13:28model sort of a privacy first instead of uh we have cheap prices but we're going to use your data
13:35for
13:35however we want right now in this data economy where everybody wants people to stop abusing their data
13:41and then using it for for everything like is it is it one of those cases where like one of
13:47the edges
13:48that you guys have not only is the environmental thing but like the idea that you have access to your
13:54data we we only store it we don't look at it we don't sell it we don't do that how
14:00does that work so so
14:01privacy first is the most important characteristic of our project so what we do is whenever we are so so
14:07that's the whole idea why we are encrypting uh the file with your own private key in your device
14:13so that means the when the file is transferred to us it's already encrypted with your private key and
14:18you need your private key to open that file so that's the idea and and we are like privacy first
14:24security is the second part third is the reliability resiliency that's that's how we have categorized it so
14:30so the outmost benefit that we offer is privacy and security of your data nobody in the entire ecosystem that
14:36storage software can ever snoop into your data or look at your data or see what's exactly residing in
14:42the data because it's always a fragment that is always lying in one single computer nobody else except
14:47the system knows where the multiple files are even if this multiple files are accessed you still need
14:52that private key which is lying on the user's desktop we don't keep any private keys or we don't keep
14:56any
14:57encryption keys with us so it's ultimately the user who controls the data and we don't do any kind of
15:03snooping on any processing by the way we are an open source project so we like the entire source
15:08code of the project is available on github people can just go and see the kind of processing that we
15:13do for data so that's this one so a power user is like i want to verify that they're telling
15:18the truth
15:18about how the project works they can just go to the source data and build yes and it's it's been
15:22audited two times and we have the audit results also placed on the earth so it's like open for everyone
15:28we just go and visit and see what's the results that's fantastic that's awesome because that's
15:33one of the things these are very complicated topics and the lack of transparency just makes them even
15:38harder to engage with doesn't it why should a normal person ordinary person care about who controls the
15:45data infrastructure because you can you've probably heard the jokes of like oh i don't care if x is you
15:52know looking at me i don't have anything i don't have any secrets i don't care i don't care why
15:57should we
15:57care so when when the cloud revolution started 20 years back that's when the hyperscalers started
16:03moving so before hyperscalers it was all like on premises and when the hyperscalers came into the
16:10market the three big hyperscalers they promised three things that would be there entirely that the
16:15cloud would provide so that seems to be the pillar of the promise why people move from on premises to
16:21the
16:21cloud so those things that like they would offer high level of security and privacy there would be
16:28less cost compared to hosting it in your in-premise on premises and the third was was unlimited scalability
16:34so when when they were hosting data in their specific location or on premises you always have a capping like
16:40if you want to get if you want to raise your uh storage capacity and compute capacity you need to
16:45do a lot of
16:46uh capex and then they to bring in servers but when you're keeping it in idc that was like very
16:50easy
16:51you could just by the click of mouse you can just upgrade your signal so the three problems out of
16:55the
16:55three promises now there's only scalability that remains so the other two was compromised over the period
17:03of time you know what happened with camera general because there were so many uh so many abuses around that
17:08so it's already it's it's it's already in public so i don't want to commend on that uh the prices
17:14also
17:14have been steadily rising uh if you see uh the cloud prices what it was promised now you see uh
17:20people
17:21end up uh like if you talk about a typical uh software organization they have like 10 to 20 percent
17:28minimum and sometimes it goes up to 30 to 40 percent of the cost the entire cost is the computing
17:33cost
17:34or the hyper skiller cost that you give so that will become a very huge cost uh that has become
17:39right now and that's that that's really something which is really concerning about your question of
17:44privacy it's it's in fact growing because of all these abuses that has happened and then there are uh
17:50some inevitable threats to the data because you are locating uh you storing your data in a public uh cloud
17:58that means it's not a private location it's like open you have the same kind of infrastructure
18:03you have the same kind of accessibility a hacker can come in and access any one device and he can
18:10get into the entire system easily because those architects everything remains the same across all
18:15the network so what happened was that that's that's a funny question that we get and in fact the one
18:21the
18:21point that you raised like i'm an SME why should i be concerned about storing my data so privately i
18:27i don't
18:27know nobody's going to be behind me but they need to understand like if there is a ransomware
18:33infection that happens if there is a virus attack that happens it generally happens with the loophole
18:38that is there in the system they're not interested in specific things so they lock in everything
18:44in respect to the fact that you have an SME you're a median size or a big multi-billion dollar
18:49organization it doesn't matter they will just come and lock everything so you need to be really very
18:54secure and you need to take that consideration into practice and of course uh because of so many of
19:00laws which are enacted everywhere because of the safety of users you really need to know where
19:04you're kind of storing your data and how secure your data is and i see that uh that uh is
19:10growing
19:11among the people right now it was not such a big uh concept or such a big threat 10 years
19:16back but
19:17now definitely people are looking in like where my state i store how secure is my data how many people
19:23can access my data it's it's really uh a concern right now and in in the era of ai also
19:29how is my
19:29data used right like is it like if if you upload it for example to to to a competitor i'm
19:36not going
19:36to name any specific competitor but they're like okay from now on we're gonna uh for this month we're
19:42not raising prices but like uh all of the stuff on your cloud is going to be used to train
19:47ai it's like
19:48well so they don't and they don't announce it publicly but uh everybody everybody knows that
19:54happens so that that's that's very uh beautiful like we we work with certain organization though
20:00those who are having a lot of proprietary data like years of organizational data and all the data
20:05related to their finance and everything they were very like they were very concerned about uh putting
20:12this uh data in the public cloud and training their elements so they came to us and they said that
20:17we
20:17want a private cloud which is not accessible to any third person and we will you set up the infrastructure
20:22for us and we will run our lms on top of that just to ensure that our lms are accessing
20:29our data and
20:29there's no third party who's accessing data at any point of time so that that's where the privacy the
20:35security of the data is really very concerning for people who are of this age right now yeah that makes
20:40sense not kind of having it out there for easy access i guess from a financial perspective why are
20:47centralized systems still dominating then because this sounds like the obvious next step but i i don't
20:54think there's there's been a move to that as high or as rapid as one would expect from such a
21:02great
21:02proposal so what happened is now as i told the hyperscalers have been in the market for past 20 years
21:09their building systems they of course they have done one of the best things that is i think the growth
21:15of cloud is really uh responsible for the growth of a lot of industry around it like the saas business
21:20only uh evolved because there are so many hyperscalers and they built in a lot of moat around the
21:25products so there's a lot of uh a lot of facilities a lot of good things that was built around
21:30storage and
21:31compute which were able to uh which the uh which the end users were able to utilize and build systems
21:38on top of
21:38that so that that that's how they've been growing and that's how they grow in the future also so
21:43that's because of the enormous benefits that they offer uh this thing and one thing is uh nobody is
21:49thrown out of job and they start hosting in any of the hyperscalers so people are like that mindset is
21:54created over 20 years of time that they're very secure of their any like scalable sort of thing uh that
22:02notion
22:02is would still be there for a long time but they they'll only be i see the decentralized cloud is
22:07going to create its own niche because of the advantage that it offers of course yeah and something
22:13i've heard a lot of times is the migration from one system to another tends to be quite the headache
22:18especially internally in a company so how hard is it to convince businesses to adopt a new
22:25infrastructure model like this one specifically okay so i'll talk about the the storage part of it
22:33the two components as you know this compute and the storage so we work only specifically
22:40so for storage uh all of them majority of them follow the s3 standard which is the amazon
22:46s3 standard that's what it is our system is combined compliant to s3 standard so you just need to
22:52change those pointers and you can really migrate your data easy so that was something that we really
22:57worked from day one we we knew that if we want to have systems built in that is easy you
23:03need to go
23:04for global standards so we have built into the system right from day one because we have adopted those
23:09standards it becomes really easy for anybody to migrate the data and use the systems that's cool
23:15and from a consumer justice perspective so you know the people themselves do you think decentralized
23:21infrastructures kind of help make digital access more equitable globally or is it kind of separate
23:28and a happy coincidence if it does uh it would be both actually uh because as i told you they
23:35took 20
23:36years to come into the stage if it had if we would have been pre-hyperscaler age where everything was
23:43on-premises and people it would have been very difficult for the growth of this whole ecosystem which
23:48is developed around so that is useful and they have done some phenomenal work but as i said there's some
23:55there are many disadvantages because of the scaling as well because of the answer why it's this kind because
24:01it's a centralized infrastructure it's easy for everyone to just attack in everyone can get access
24:06because it's all standardized you just need to access one location and you can get access to a lot
24:10of locations because of the same architecture same design and thing so that's that's how the decentralized
24:16system comes uh as a separate entity where we can have niche around build around that products
24:22where we can scale because the kind of scalability that we offer and the security that we offer it's very
24:27that them to do that at the same way because we use blockchain nodes and all those kind of things
24:33so it's a complex architecture it's it's like a very uh uh it's like a lot of things built into
24:39a
24:39h and a fit onto else so it works in a perfect architecture no that makes sense and i guess
24:45in
24:46the climate space do you think decentralization is genuinely a solution in terms of energy use and
24:52water use and what we were talking about with storage centers being different to computing centers
24:56or is it risking shifting energy use somewhere else no it's definitely so because i i told you
25:03because the data centers is one of the highest consumers of electricity energy all over the world
25:08you can hear if there's a new data center set up in some town county they have come up with
25:13so much
25:13of opposition that they had to shift the plant altogether there were so many instances around that
25:20so definitely it has a lot of the decentralized cloud does have offer a lot of advantage not in
25:27terms of you under utilizing the underutilized idc and saving those kind of resources which is like
25:33goes very high because we are saving so much of this fuels green gas emissions and all because we are
25:40utilizing the underutilized and it becomes benefit for them as well because you have an infrastructure set up and you
25:46have
25:4620 percent of the infrastructure 30 percent of the infrastructure just like ideal and you need run all
25:52those infrastructure because you need to have you the system has to operate at 100 percent so if you
25:57save multiple geographies around the world and you save so much players in this it the the effect is
26:03compounding and it's always uh beneficial and then i would imagine that it would it would help keep
26:09keep let's say closer to not obsolete but closer to obsolete data centers alive and not have uh all of
26:17the because there's a there's more to data centers right there's also like if something becomes a little
26:23bit more obsolete it goes into the lambs landfill if it belongs to something cutting edge right um he tells
26:30a little bit more about the environmental side of this right like the the carbon emissions and stuff like
26:36that and how like is is um all of the efficiency gains basically from finding those uh those
26:44underutilized data centers and working with them or is there more that can be done to mitigate uh to
26:51mitigate power consumption and such i think uh i think uh people uh uh the organization can look at uh
26:59using
26:59decentralized cloud storage in a better way so they can they can shift in some load from the centralized agency
27:05but where it's possible and it helps so we can grow a system and we can get in more idc's
27:11which are
27:12under utilizing energies and all those things so that will definitely help to reduce the garbage footprints
27:17around the world i can save globally like you can save so much amount of energy for being wasted around
27:22this
27:22is there a difference for example i actually i actually don't know like is there a difference depending
27:28on what type of storage is used on the data centers right like uh uh i would imagine like how
27:34how do you
27:35how do these data centers operate in the modern day do they even use spinning disk do they all use
27:39a
27:40solid state like how does it work is it a hybrid so it's a hybrid system right now so they
27:46are
27:47because there's so many of the circulating handles it becomes very difficult for them but they are being
27:52phased out to SSDs and then we have now more efficient NVMe so those are the ones which are coming
27:58up
27:58but even to phase this out it would be it would take a long time because these developments that has
28:03happened like the NVMe the SSD will just recently compared to the spinning disks which were there for
28:09so many long years and now they are no more produced but but the kind of efficiency that is happening
28:14is
28:14really very good so we feel that this this would all be recycled very soon and then we would have
28:19more
28:19efficient less energy consuming kind of thing because what happens is when this disks which are like
28:25cylindrical disks they are rotating they are generating heat and all those things so that becomes a real
28:30concern in terms of environment effect and the carbon emissions and all those things but when we have
28:35the solid disks when we have SSDs when we have NVMe the heat is reduced very late and because there
28:42is no
28:43physical rotation happening those kind of costs can also be saved over the period of time so in order this
28:49becomes much more efficient this becomes smaller in size when when the device becomes smaller inside they need
28:54less cooling they need lesser land land to be cooled so they need less power so all those factors play
29:01a
29:02lot of role in ensuring that the carbon emission goes down and the energy efficiency is increased
29:09why why are data centers underutilized is it like a pride thing a lot of the companies want their own
29:15data
29:16centers so they're just going to build more instead of uh finding deals to like utilize uh the underutilized
29:22one over here and over here and over here yes so so what happens is i'm talking about this in
29:28uh the uh
29:29whenever the data center is set up they'll set up for the period of like 10 years because it's a
29:34huge
29:34investment to set up the data center they can't set it up like just so if if there is uh
29:39there is an
29:40energy utilization that they think of like 100 racks they would end up uh setting up the infrastructure
29:46would be like 500 racks that is the standard when they set it up so because they set it up
29:51such a
29:52way it's always and what happens is when you go suppose you have five floors of data centers you have
29:58one operating at full capacity the moment you go to 101 you will have to take one more floor and
30:04then you
30:04will have to write down the entire infrastructure because that would be big cost which is important so
30:09that's something that is that that is still going on but the new technologies which are coming in that will
30:15ensure that they like like built-in rack which is cooling so you don't have to cool the entire flow
30:20you just have to cool the entire rack itself and then there are of course new things coming around
30:24oceans and the space idc which are coming up so all these things will really go a long way to
30:30ensure
30:30that we have those kind of energy efficiency build up in the data centers that's super cool i guess i
30:37want to close out with what are you most excited about in this future of decentralized tech what are you
30:42most
30:42looking forward to i i think i think the benefits that we offer we are able to offer the clients
30:47because of this decentralization is really some very good benefit and because of the carbon emissions
30:53and the global footprints of carbon that we save that we are really very excited and committed in
31:00this mission of decentralized cloud storage amazing well guys you can check out the blog at
31:05tangeliclife.org tangelictalks to learn more about storix network and murphy john as well um is there
31:14any closing words guys anything you want to leave the audience with oh no i think it's a great time
31:19to be
31:19alive and we want to be uh we utilize whatever is what are the new technologies coming up i think
31:25it's
31:26it's great to be the part of it i love it amazing well we'll catch you guys on the next
31:30one thank you
31:38let's talk power let's talk change for rural lights to brighter days equity rising voices strong we're
31:48building tomorrow where we all belong tangelictalks energy equity pride empowering the world side by
31:58side a spark becomes a fire a vision that's true together we rise it starts with you
32:12we rise it starts with you
Comments