Player is loading...

Embed

Embed code

Transcriptions

Note: this content has been automatically generated.
00:00:00
that's the last one i thank you very much for um
00:00:04
hello everybody i would like to talk with you today about uh how
00:00:09
the container or doctor revolutionised the that ops i. t. landscape
00:00:14
and my talk will not be necessarily very technically i mean it would
00:00:20
be about doctor it'll be technically but without code examples because
00:00:23
i wanna show you or explain you more the principle of the
00:00:27
core concepts behind doctor and also the results of these concepts
00:00:32
that will change your uh they leave work if you if you working with doctor
00:00:38
um before we start that introduce myself my name is about him
00:00:42
oh i uh work for the company called eight years and if you would like to
00:00:45
find more about me you can best is to for me on the top
00:00:50
and or into it so then you can see my reporter isn't it see what i'm doing and we should was doesn't start starring
00:00:58
uh yeah i'm working for the company in case and we actually developing software
00:01:04
and this also today involves developing um clout infrastructures
00:01:10
because developing software and developing infrastructure is actually the same it's both software
00:01:17
on one hand it's softer that produces a code producer website application
00:01:21
on one hand you develop software produces an infrastructure which is also actually saying
00:01:26
you have to test it you have to have what you sure
00:01:28
once you have a have a repository of the continues to ration can you do that very and all those up a sort of stuff
00:01:36
then we also do some workshops in an area of docket
00:01:40
continues rationed and holding company does the clustering infrastructure
00:01:44
but at dive into doctor who viewed no was what grace
00:01:52
who of you used already doctor
00:01:56
what he already
00:01:59
so but nevertheless well what is the idea behind talk
00:02:04
or the principal idea behind doctor is to
00:02:07
have an application which is isolated complete that runs
00:02:13
everywhere isolated means that there are no conflicts
00:02:18
or whatever to other libraries that are on the system so there is no conflict
00:02:23
there are no dependencies to other libraries that are uh on the system so you're know
00:02:27
exactly what you're running complete means that the application that you're running brings everything
00:02:33
with it it it needs so everything that application needs is part of the the package
00:02:41
so and running everywhere means that the result or the container is really
00:02:47
running everywhere i mean everywhere means in on linux okay so
00:02:54
the other everything else doesn't matter no seriously and that means is wrought everywhere and and
00:03:01
and and i mean now it's basically only looks so it runs on everything system
00:03:07
yes some button colour parameters to talk right um but also this
00:03:11
running on windows since uh a couple of weeks enough
00:03:15
windows supporting a doctor already so we can really run native
00:03:19
windows applications on on the on the windows kernel
00:03:23
oh and uh we probably one day they will be also something with the mac but um
00:03:29
but why is the possible but i mean doctor is two years old you know
00:03:34
of that and why is it so popular i mean we we had a uh
00:03:38
the lars also it's called even solar containers that exists is two thousand five
00:03:44
in a very successful and and and people on as of our as far
00:03:48
as a systems uh use it we have a freebies the jails
00:03:52
it exist since the year two thousand or so we we have this container concepts um
00:03:58
in in the history of of of a unix is all the time to my
00:04:03
is is is doctor so successful in that in recent years and it's actually
00:04:08
the focus of doctor specifically to be user friendly to be easy to use
00:04:14
the city for developers and also with functionality to be very
00:04:19
easy extendable so you can really easy extends talker and
00:04:24
what new stuff in and you can extend with new functionality very easy much much easier than the other solutions
00:04:31
but understands advantage let's take a look at uh on how
00:04:34
doctor is is 'cause conception list you know from
00:04:40
the object oriented software development from the object oriented languages that there is something like an object
00:04:46
in the class and an object is nothing more than an instance of the class
00:04:51
but the same you can think of of images and containers in a doctor minor you have
00:04:55
an image which is represents class and if you instantiated make it run it becomes
00:05:02
a container because it's it's the actual instance running so a container is the white you you specific
00:05:08
my uh you specific drop for example or um yet so different instances of of this image
00:05:18
so
00:05:20
but each image itself contains of multiple layers on those layers can
00:05:26
be stacked together on each other on top of each other
00:05:29
so actually if you would take a look at what press image
00:05:33
it's not just the purpose image itself it contains a
00:05:37
underneath it contains p. l. p. h. p. layer that provides your peach
00:05:41
be a run time environment an underneath it it contains an uh
00:05:46
and stripped down the ins operating system in this case it will be unix or so are you going to
00:05:54
and so this is how you actually build your applications it's like a lady layer by layer
00:06:00
or image that extent other images and you can read mail more more complicated cushions
00:06:09
um so
00:06:12
when we talk about talk or we actually mean various different things so we don't we actually need things like
00:06:20
the doctor demon which is responsible for starting the containers popping the containers
00:06:26
and running two containers but also is responsible to get the
00:06:28
images to build the images and we talk about soccer uh and
00:06:33
we need to the client like the command line client
00:06:36
yeah that is communicating with the demon and is responsible to trigger the demon
00:06:41
to tell it may start start the container and stop the container and
00:06:45
greater network whatever and we have also the the registry which is not part of
00:06:50
of the great self but is very uh tightly coupled to doctor because it's
00:06:55
provides a repository where the images that i create or someone
00:06:59
else creates i can make any use for my purposes
00:07:05
and
00:07:08
then next to the doctor functionality we have the the the common extensions
00:07:14
most of them are all this this example see uh provided
00:07:17
by right doctor incorporated itself but they're very closely coupled to doctor
00:07:22
and we're talking about uh composed local compose which is uh
00:07:26
allows you to create an application uh that is just of multiple containers we kind
00:07:32
of a complete compose multiple containers together to run them as a whole
00:07:38
and we have a machine doctor machine that isolates you to create
00:07:43
what the provision service so you can remote use machine do provisions server where
00:07:48
doctor will be installed and it's configures you can communicate easily with it
00:07:52
and swarm is actually responsible for if you have a multiple doctor votes
00:07:57
um but i can go to communicate with each others you can create a cluster of
00:08:01
multiple other instances and this is what is you know we use swarm to use
00:08:09
oh
00:08:12
but one of them um main reason why doctor is
00:08:16
so successful is that the eco system around doctor
00:08:22
is providing hundreds and thousands of extensions or product that does specifically made for doctor
00:08:30
it extends the functionality of taco and provide new functionality that
00:08:35
would really um let me talk just on itself um
00:08:42
let's look of um a group of those tools that exist in the it was system so
00:08:48
we we need to better understand them we need to group them into into categories
00:08:53
and we have a good grief tools which is a which are responsible for august ration
00:08:58
august ration means if i have multiple containers on us or so in the containers if i have multiple
00:09:03
hosts a cloud environment okay i've on the machines and that won't deploy on those hundred machines containers
00:09:11
very easily those hundred machines need to communicate in some windows when critic another cluster many to orchestrate
00:09:18
my containers i need to orchestrate a witty containers executed actually and
00:09:23
so that tools like when it is like measles likes war
00:09:26
and or ranger that helped or for c. v. it the the orchestration of of containers across multiple
00:09:32
holes not multiple environments and then we have tools that are responsible for monitoring you containers
00:09:39
and logging so that you always know was no matter where it continues running that is correctly
00:09:44
the the data is locked correctly and that the state of the containers money toward
00:09:49
and then as to see their response before discovery and discovery is needed
00:09:54
if you have a a environment way of like hundred containers
00:09:58
michael services running and working together those containers need to communicate
00:10:03
to each other those continuously to find each other
00:10:06
'cause you can not just a run the container one node and run on on something on on the other note
00:10:12
uh and expect that they find each other and they can communicate with each other so you need tools
00:10:16
that facility with this and it example of those tools are for example onto the t. c. d.
00:10:23
as we keep our net flicks a rica that help you to uh
00:10:28
uh to achieve this goal very easily in context of doctor and then you have also tools that are
00:10:34
of course wages virtual networks we have a network inside a network so it you containers can communicate with each other
00:10:41
like it was a a bunch of individual network and there's also plenty of tools in this area and well
00:10:52
let's look before we go to next like let's look on the on the um
00:10:59
on the architecture of of of a doctor so we have here um
00:11:06
our and architecture of uh of taco so we have on the need we have the hardware and
00:11:11
then we have some linux stuff going on here so there's like the linux kernel functionalities
00:11:16
uh like names basis you so there's really something deep inside inside unix and then on
00:11:22
on each on the linux system we have the container running as i told you that the containers everything that's needs to run
00:11:30
it contains all those leaves libraries then you get for your application to run like the g. d.
00:11:34
n. like no g. s. like ruby libraries everything is there in the container and the demon
00:11:41
or actually i called a a rota doctor but is actually the
00:11:44
demon demon is not in the container it's just communicating
00:11:48
it's not communicating with the container itself or not directly it's it's
00:11:52
starting this containers managing those containers it sees if they
00:11:56
ever running into stops them and so it's kind of a sits inside of the containers not part of it and
00:12:04
this this allows uh some third party to coming to communicate with
00:12:08
the demon that that will then start and stop the containers
00:12:13
and one thing to note is that the actually the container all running on the one on the same colonel
00:12:19
you don't have a multiple kernels yeah you have one hardware one
00:12:23
unix kernel and all those containers are running on one kernel
00:12:28
and it hasn't very um big impact on performance opposed
00:12:33
it in this case because everything's running on a
00:12:37
on one single car isolated from each other it takes
00:12:42
zero ahead regarding the start the container it doesn't matter if you start your my s.
00:12:49
q. l. or or p. h. p. application without containers with containers it's zero overhead
00:12:55
because actually this is the same curl and the same ruffle here and
00:13:00
the second part is um because it's it's isolated another sorry not eyes
00:13:06
that because it's running on the colonel resource overhead is zero
00:13:10
or like a like an virtual machines where you haven't operating system and
00:13:15
operating system and operating system on one hardware where it's come zooms
00:13:19
and and and it's your resources here you have absolutely no resource
00:13:24
overhead so it does not consume more resources than single process
00:13:30
well that's
00:13:33
it'll why this this this dog architecture is is such a revolution or
00:13:37
such a big change in in how uh you develop your applications
00:13:44
well
00:13:46
before we die that is there are some things that really um intimate doctor
00:13:53
narrows down or so or restricts you when using so there is a case where
00:14:00
we're um document is ideal for and that's the case where you have um status containers so
00:14:07
the requirement for doctor is that the container status does not any state
00:14:12
or you create a container that has an external state so we can have something in that connect to the database
00:14:17
where the state is inside but your application itself is is status no we're talking here about i don't
00:14:23
know if if you know the twelve factor applications well factor apps the ring a bell so you should do
00:14:31
with twelve factor apps it's it's a it's a design principle how you develop application of the status and
00:14:39
because this restriction is in many cases um yeah oh well a lot of people this is um
00:14:47
the same him at the beginning it's very problematic it's very difficult to design and application
00:14:53
especially if you haven't exist application it's very very difficult to migrate or to move it into the state this one because you
00:15:00
have a lot of overhead you have a lot of uh things already in your application and it's very difficult to
00:15:06
uh to to migrate to make it status in many cases noble but it's only it's it's it's it's problem
00:15:14
and um this this um restriction is on the other hand that you'd huge advantage
00:15:23
and a huge advantage is that you have this so called atomic rule we containers so you have that in container
00:15:30
it's actually a thruway container this means that you start the container and if you want to stop it
00:15:37
you don't care you just stop it you remove it and it's gone so there is
00:15:44
no old libraries left anymore so there's nothing on your host operating system
00:15:48
but uh maybe leftovers the next time you start it doesn't work anymore in
00:15:52
if you update the next version it doesn't work anymore this is
00:15:55
uh not a problem anymore there's really no leftovers because it has no state you can just stop it you don't have to worry
00:16:02
and because it is posted you can start it because it has no state you can start it once twice three times you can run it on
00:16:10
one data centre you can write on two data centres you can run all over the globe
00:16:14
has has no state it's easy to scale up this huge advantage if you wanna have a publication it's not
00:16:20
only a scalable so is that you know can cope with millions of users but also that it was
00:16:28
also the application it's really resigned and this means you can start it and if it stops you started again
00:16:34
where you started on a new server which is a reliable so it's really improved the real real
00:16:42
the silence of implication worried but um and this is of course the result of the immutable container so we call it uh
00:16:51
the true we continue recording mutable that because it has no state and so this
00:17:01
so if we take a look or if you if you if you
00:17:04
using doctor for a while you know then you will uh ask yourself maybe you know how to stop actually
00:17:10
compared to other technologies you know how does it compare to the m. there because it's some sort
00:17:15
of a visualisation you know but how does it really compared to the empire anna how does maybe compared
00:17:21
unstable because you could you can isn't integration with it and if you if you if you
00:17:26
uh like doctor versus something on the internet you will find many articles and many results
00:17:32
how a doctor is compared to all those technologies but if you take a bit to close the
00:17:38
look on the slide you maybe see if you know this technologies you maybe see that
00:17:42
at that regroup the technologies we're talking about the we have a bunch of asian technology like
00:17:48
v. and we're like send key v. m. but is the vegetation technology we have
00:17:54
the figure ration management tools you know like uh and civil like soul stick cheek up strong now
00:17:59
those out configuration tools and also we have some who was like open steak nails done
00:18:05
more for orchestration and you see that actually doctors some somewhere in between so it
00:18:11
is we compares to things like the empire and also to the sensible but that compares
00:18:18
to uh to to mean different things so that the result is actually that
00:18:24
the documentation is not very focusing on on on the replay six ecology but it's a very
00:18:33
uh brodie range not very deeply itself but very broadly across multiple technologies
00:18:40
'cause its cover something from configuration management cover something for vegetation
00:18:44
it's not good in but relation compared to the younger or something because
00:18:48
we um has really really great functionality regarding visualisation but it has
00:18:54
something the core functionality of it is there and those many factors
00:19:01
of of soccer because it spends all but this much technologies
00:19:05
allows a completely different way how how people can read incorporate together all the break together because
00:19:14
this vied orientation that new world falls possible that are not possible otherwise and i mean
00:19:22
ruffles for for developers for operations teams for just about as big as for for work you
00:19:28
ace but um would not be possible otherwise well before we answer the questions well
00:19:37
what kind of what falls let's see what a typical work full day would be without uh
00:19:44
having talker so in a classical after the deployment software development world we have this
00:19:52
kind of work full so let me explain we have two teams here are responsible for
00:19:59
uh let's say developing and deploying application into
00:20:03
production making available fall for the customers
00:20:07
to visit the website for example we have on one hand we have developers and say okay i have application
00:20:13
i wanna have it running somewhere and have it the the ops people on the other
00:20:17
side just wonder straighter that are responsible for providing you the infrastructure uh so yup
00:20:23
occasion can run and so you you make you press okay i need hardware i hit
00:20:28
machine that has thirtieth uh thirty thirty two gigabyte of ram i have must
00:20:32
need machine has to uh as as the drives with the this i i ups rate
00:20:36
and stuff like this and then you get the machine one day when you
00:20:41
start configuring it because maybe it is already configured but anyway even if it's configured
00:20:47
you still need to do a configuration because you need to adopt a machine
00:20:50
or you specific purposes you need to install software you configure it and then you say okay i configured everything
00:20:57
and then there's someone from the of ops team or operations team looking in it and say we will
00:21:03
know is not allowed an outcome organisation and then it you start you start tweaking around so you start
00:21:10
looking around and see what is allowed how can i work around and if it's wide larger organisation
00:21:16
it's it's very strictly because you cannot install days you get here you have to use only this operating system because those
00:21:22
people wanna have also an easy life you know they want they don't wanna support like twenty five different operating systems
00:21:27
it would have one operating system you want to or a centre was or baby and they don't wanna like twenty
00:21:33
five different because you know i'm a baby and it's easy for you but then it's easier to maintain one
00:21:41
operating system it easier to for them to maintain only one set of libraries you know it's easy for them to
00:21:46
manage one j. v. m. instead of twenty so and that it's it's starting triggering and it it goes
00:21:53
on and on and on until one day everyone is happy and your applications find a running and then browse is
00:22:02
that and this is a very cumbersome up the approach because it takes a long time it's very hot
00:22:08
establish of work flow inside an organisation where different teams liberate with each other on the topic they're not very
00:22:14
familiar with because developers are not very familiar about the
00:22:18
infrastructure because they know that something like this exist
00:22:22
i didn't know exactly what it is and how to cope with this on the other hand the
00:22:26
operations people you don't know the application they don't know what is inside and how to
00:22:31
how to run and with the doctor approach or container push it's it's very easy no you
00:22:38
have the developer that graces application yet the developer that rates a container where his application
00:22:45
is running you tested it verifies it makes sure that his application is running this container and
00:22:51
is running in this container successfully after that it just ends or the container but it
00:22:56
helps people that operation people say look this containers running i give it to you make it
00:23:02
run for me and the present team just figures everything around the parameters around it
00:23:10
and can easily run the application because it doesn't have to know what is actually inside just has no
00:23:16
how to start the container it doesn't have to know how to start application how to connect
00:23:21
to the database how to do this how to do that that's really application specific stuff
00:23:25
it doesn't have to know what it just has no i'll have right should i continue my environment to make this
00:23:31
a container around and because because of this this rightful we
00:23:39
can say that doctor is creating a standardised interface
00:23:44
not only for the operating system and the container to communicate with
00:23:48
each other but it rains also a standardised interface between
00:23:52
a developer someone hand and operations people on the other hand how they communicate because if they could
00:23:59
we could each other they have clear interfaces and how they communicate because they always communicating
00:24:05
around the functionality docket provides and there is no mismatch and this is called the location of duties so we have
00:24:16
a really strict segregation if between the developers and the
00:24:20
operations and let's see it from a different perspective
00:24:24
uh on this topic so um we have here at classical
00:24:29
layout of uh our um infrastructure on the left
00:24:33
hand side we see um the classical view so we have the infrastructure hardware we haven't operating system
00:24:41
where the hyper wiser and and uh oh each type of wiser we have in the two machine running the end there whatever
00:24:47
and inside the stadium where we have an operating system and its operating system contains some libraries and then
00:24:52
with the separation and this is the same this is the same so we have actually
00:24:57
one two three four operating system that has to be manage on the other hand we have
00:25:03
this the doctor set up we have here infrastructure we haven't operating system we have hated
00:25:08
upper engine and on top of this we have the containers which actually like an operating system
00:25:15
because it's a doctor engine takes care of it and this is kind of you
00:25:19
know approach and and this this interfaces come into place because this and this other domain of
00:25:27
the operations people is that the domain but operations people operate a crate infrastructure the
00:25:33
maintain infrastructure to make sure that those things around uh yeah yeah it's this part and
00:25:39
here it's it's a bit less and this is the upcoming and on the other
00:25:45
hand we have here the developer domain and here the developer to the u. c. developer domain
00:25:54
and up to maine completely separated because here in classical what you as a developer yes
00:26:00
you have to know what is actually running on the operating system that is not under
00:26:05
your control you maybe have access to it but you have no control over it
00:26:11
at least in in in a large organisation we have we have two teams and this results
00:26:18
in conflicts this is the area way he has to communicate with the other side
00:26:23
without the language and the interface you have to communicate here with the they have and
00:26:31
and the all the the operations on on the same stuff without technically interfaces and
00:26:38
the the ops have a different view on this then the developers and this is normal and
00:26:43
here it's very clear there's that the separation between them too so finally we can say
00:26:51
that if we talk about containers we can say uh that the
00:26:56
developers are responsible for making the container run them responsible for
00:27:02
making the application in the container run well and operations people
00:27:07
are responsible will actually making the container run somewhere that
00:27:12
is uh the final thought about this so if we talking about
00:27:16
desegregation of duties between the deaths and the ops that's
00:27:21
still need something like that pops it is a really hyped worth you know and everyone is talking about that pops
00:27:27
and i'm telling you that we don't need it anymore i'm not telling but i'm asking the question do
00:27:31
we need any more but you know before we dive into this does that ops buzzword topic
00:27:36
let's find out you know what does this devil stuff mean no if we talk about their tops
00:27:44
and the definition that different things meant by this you know that pops or deaths doing up
00:27:52
stuff is actually what your your company things you have ups means you know because they think okay we
00:27:58
don't need also we have that ops they do everything is what but that's not quite true
00:28:03
now but then the second definition of their hopes is actually that means that uh you're developing or
00:28:14
someone also somewhere is developing the infrastructure just like
00:28:19
it's normal application just as normal code
00:28:22
it has a it has a builds estimate is that the u. i. system it is
00:28:26
a continues integration it has continued delivery is just nothing more than a normal
00:28:30
application and you have to develop it just a normal application and this results that
00:28:38
uh you're a dead hopes use actually the same tools as a developer
00:28:42
you have an i. d. you have a a continues tests and everything like this
00:28:48
and if the definition which is actually not a definition on the technology but
00:28:53
it's a definition on that that helps means that developers and operations timit closely
00:28:59
collaborating together as one t. not as a separate fumes of is actually
00:29:04
you can say it not as one person but as one people in that uh what works closely
00:29:10
together so developers learn from the ops what are they doing and know what are they doing
00:29:15
and their domain and on the other hand as well that ops normal to developers
00:29:19
are doing because they're operating actually in the same domain now ended operating in
00:29:30
separate separate domains the it together because we can assume that
00:29:38
the deaths and all use the same tools to get
00:29:42
work done and because they're using the same tools everyone from each side knows how to use those tools
00:29:49
and this is for example doctor is such a tool because it's used on one hand it used by that but all people
00:29:56
on the other hand it use also by the deaf people and to to highlight this example
00:30:02
let's take a look at the the uh the doctor run time environment so well
00:30:07
we have here a run time environment uh it's it's called the doctor
00:30:14
away container environment so we we have a container that actually
00:30:18
knows that it's running in the court that it's container it's knows that it's running on the docket demon because it coming in
00:30:24
can it can communicate with the doctor engine or the curry demon
00:30:28
and this is a great example for this for deaf son
00:30:31
ops using the same tools because this is our application it is
00:30:36
responsible or it is maintained and developed by the developers
00:30:40
because it contains your business data and business logic and this
00:30:43
is your business application that's developed and is uh containers
00:30:47
but i actually developed by the ops people and those containers are responsible for security
00:30:53
maybe securing the application or securing the whole infrastructure we don't know but
00:30:58
they use the same tools to achieve this goal they use a logging but you the same goals because
00:31:04
they're using using containers that are aware of of of soccer and that can maybe lock your application
00:31:10
without you even knowing it but it yeah actually i have not uh who was one tool that
00:31:18
use uh from both sides in a in a different context in it in a different manner
00:31:24
and the standards that interfaces cannot only be used from from that and obsolete also
00:31:30
is within the containers together so the containers can communicate with each other
00:31:35
and use information from one container in the other container
00:31:38
without having interfaces we help without i'm having
00:31:44
without being tightly coupled to each other so we have something like locking we can provide very easily looking for container that
00:31:51
is running someone a server that aggregates everything makes a really
00:31:54
pretty lot across you all infrastructure without the application
00:31:58
or without the continue even noticing that is log be locked and the same goes for archiving we don't have to
00:32:04
implement some archiving functionality normalisation which is no because it has
00:32:09
some interfaces and doctor how to archive docket containers
00:32:12
someone would do it for us and application doesn't have to care about this and this goes on for security for
00:32:19
august rating multiple containers to sell what what are the tools of for the uh the upstream
00:32:23
of the containers and the containers don't know what about it they just know that that
00:32:27
i don't even know that the containers that think they're running it in they all operating system and you can yeah is this goes in all
00:32:34
on an on so the summary we can say that even though well
00:32:44
doctor is a very very popular two or two in the recent
00:32:49
recent years on a um there's one yeah difficulty actually is that doctors often
00:32:58
used more by developers then by operations and it has some reasons
00:33:05
for that um because one of the reasons is that um operations people
00:33:12
ah used to use their tools they know they're the best
00:33:17
shells they annoy their shells in all very precisely how to configure linux they
00:33:22
know all this nifty details of linux kernel you know everything and
00:33:27
now with doctor never a big limitation in this is flexibility disability and
00:33:34
they have to to learn new tools they have to learn
00:33:37
uh to use new new functionality and this means that um the operations
00:33:43
people need to change how the working and a lot of those
00:33:48
people don't like change i mean a lot of actually all of all of us don't like changes if it's if it's too much
00:33:55
um then you have no control over the containers so if someone is building a container if it's you when you
00:34:02
have control of you container that someone else operations on of the controller would they don't know what it's inside
00:34:08
and for for many people this is a a big problem because if they don't know what's inside
00:34:13
how can they operated how can they make it run so this is uh
00:34:17
mm yeah this is a full thought of the operations environments it's
00:34:23
it's a problem uh or how many people have a problem with that
00:34:29
so so the conclusion can say that the containers provided better operation
00:34:37
of responsibilities between the operations and developers and that we have
00:34:44
or we have to take care that our container
00:34:47
that we create a status because if they don't statements we can we can we can make them run
00:34:53
but they will not run efficiently or the the waves problems with running and we have the possibility that
00:35:01
doctor on itself is um it's just an application better
00:35:06
decision so we which allies in our application
00:35:09
nothing more with some tools around it but in principle it's it's it's a very
00:35:15
limited functionality it's it's white inspectors limited and if we wanna use doctor efficiently
00:35:22
we have to rely on that party applications and third party tools and make use of them create
00:35:27
better locations especially if we're a moving into microsoft's
00:35:32
architecture where we where we have multiple containers
00:35:35
communicating with each other for multiple hosts then the drawback is that we have to learn new tools
00:35:43
i mean we as a developer at after soccer and all those talk a specific tools
00:35:49
and of course the operations people need to uh to learn those tools um on the other
00:35:55
hand the um the place i think few years old uh two years or something oh
00:36:01
it's it's it's very uh the first release was two years ago two and half years but
00:36:05
i think the the first public release of doctor was two dozen thirty the first
00:36:10
a a one zero releasable was to doesn't fourteen or something i'm not
00:36:15
sure but only a few years and we have a thousand hundreds
00:36:22
and thousands of of technologies around occur and how you can extend doctor
00:36:28
and how you need to extend docket to use it efficiently
00:36:31
and this is sometimes very very confusing are very difficult to pick the right tool i mean it's
00:36:35
difficult to find the right to and then if you find if you want is very difficult
00:36:40
pick the right to for your specific use case because all those tools that exist exists for
00:36:45
for for reason because they're different and it's it's sometimes it's very difficult to make a
00:36:50
a tool and also to stick with it because i know a lot of those tools
00:36:54
i going extinct in in months years and uh you don't wanna you know that
00:37:01
on i'm not from morse and last but not least there's that still up not yeah some
00:37:07
kind of a problem is that a lot of stuff the vendors not provide container
00:37:13
yet i mean there is something like a official containers in that in that document just
00:37:17
re joe called official but i mean they're not really official they're just officially him
00:37:23
managed by doctor incorporated but those are not official containers by the companies who rate is software
00:37:30
you know and off the we have software and operates to which is open source because
00:37:35
you can download you can start it there's not they're not many properties software i mean there is
00:37:41
no oracle database in in in the registry and not so many a commercially not open
00:37:47
source of the uh on on on the registry not yet at least and so this is
00:37:53
a bit of a problem for for at least a large corporations lotta conversations to
00:37:58
adopt doctor because they fear that there might be a um a problems yeah and that's it
00:38:07
for my talk i hope you have a good insight about the consequences of doctor and
00:38:13
how it will change your life or life if if you're working with it and how

Share this talk: 


Conference program

Keynote
Jean-Baptiste Clion, Coordinator DevFest Switzerland
26 Nov. 2016 · 9:40 a.m.
197 views
How to convince organization to adopt a new technology
Daria Mühlethaler, Swisscom / Zürich, Switzerland
26 Nov. 2016 · 10:14 a.m.
116 views
Q&A - How to convince organization to adopt a new technology
Daria Mühlethaler, Swisscom / Zürich, Switzerland
26 Nov. 2016 · 10:38 a.m.
Animations for a better user experience
Lorica Claesson, Nordic Usability / Zürich, Switzerland
26 Nov. 2016 · 11:01 a.m.
121 views
Q&A - Animations for a better user experience
Lorica Claesson, Nordic Usability / Zürich, Switzerland
26 Nov. 2016 · 11:27 a.m.
Artificial Intelligence at Swisscom
Andreea Hossmann, Swisscom / Bern, Switzerland
26 Nov. 2016 · 1:01 p.m.
193 views
Q&A - Artificial Intelligence at Swisscom
Andreea Hossmann, Swisscom / Bern, Switzerland
26 Nov. 2016 · 1:29 p.m.
An introduction to TensorFlow
Mihaela Rosca, Google / London, England
26 Nov. 2016 · 2:01 p.m.
305 views
Q&A - An introduction to TensorFlow
Mihaela Rosca, Google
26 Nov. 2016 · 2:35 p.m.
Limbic system using Tensorflow
Gema Parreño Piqueras, Tetuan Valley / Madrid, Spain
26 Nov. 2016 · 3:31 p.m.
331 views
Q&A - Limbic system using Tensorflow
Gema Parreño Piqueras, Tetuan Valley / Madrid, Spain
26 Nov. 2016 · 4:04 p.m.
How Docker revolutionized the IT landscape
Vadim Bauer, 8gears AG / Zürich, Switzerland
26 Nov. 2016 · 4:32 p.m.
Closing Remarks
Jacques Supcik, Professeur, Filière Télécommunications, Institut iSIS, HEFr
26 Nov. 2016 · 5:11 p.m.
Rosie: clean use case framework
Jorge Barroso, Karumi / Madrid, Spain
27 Nov. 2016 · 10:05 a.m.
Q&A - Rosie: clean use case framework
Jorge Barroso, Karumi / Madrid, Spain
27 Nov. 2016 · 10:39 a.m.
The Firebase tier for your app
Matteo Bonifazi, Technogym / Cesena, Italy
27 Nov. 2016 · 10:49 a.m.
Q&A - The Firebase tier for your app
Matteo Bonifazi, Technogym / Cesena, Italy
27 Nov. 2016 · 11:32 a.m.
PERFMATTERS for Android
Hasan Hosgel, ImmobilienScout24 / Berlin, Germany
27 Nov. 2016 · 11:45 a.m.
Q&A - PERFMATTERS for Android
Hasan Hosgel, ImmobilienScout24 / Berlin, Germany
27 Nov. 2016 · 12:22 p.m.
Managing your online presence on Google Search
John Mueller, Google / Zürich, Switzerland
27 Nov. 2016 · 1:29 p.m.
Q&A - Managing your online presence on Google Search
John Mueller, Google / Zürich, Switzerland
27 Nov. 2016 · 2:02 p.m.
Design for Conversation
Henrik Vendelbo, The Digital Gap / Zurich, Switzerland
27 Nov. 2016 · 2:30 p.m.
Q&A - Design for Conversation
Henrik Vendelbo, The Digital Gap / Zurich, Switzerland
27 Nov. 2016 · 3:09 p.m.
Firebase with Angular 2 - the perfect match
Christoffer Noring, OVO Energy / London, England
27 Nov. 2016 · 4:05 p.m.
Q&A - Firebase with Angular 2 - the perfect match
Christoffer Noring, OVO Energy / London, England
27 Nov. 2016 · 4:33 p.m.
Wanna more fire? - Let's try polymerfire!
Sofiya Huts, JustAnswer / Lviv, Ukraine
27 Nov. 2016 · 5 p.m.
Q&A - Wanna more fire? - Let's try polymerfire!
Sofiya Huts, JustAnswer / Lviv, Ukraine
27 Nov. 2016 · 5:38 p.m.
Closing Remarks
Panel
27 Nov. 2016 · 5:44 p.m.

Recommended talks

SIGCHI Social Impact Award Talk
Leysia Palen, University of Colorado Boulder, Boulder, United States
20 April 2015 · 2:34 p.m.
852 views