Designing Blueprints for Cloud Applications

Diposting oleh Unknown | 17.50

hello i'm katherine word from Cisco's cloud and Systems Management Technology
Group
here to present the latest innovation within our intelligent automation for
cloud suite
which is called the application stack accelerator pack
but which we refer to as pancake because it enables you to create stacks
it's probably best if you have a basic understanding of Cisco intelligent
automation for cloud before viewing this video
you can get that by going to www
that Cisco dot com flash go slash
I a cloud cisco has developed cloud accelerators are content cartridges
that snapped directly into the Cisco intelligent automation for cloud
management platform
this video series covers the application stack accelerator
once a successful writer is installed consumers can configure and consume
multi-tier applications
from the Cisco intelligent automation for cloud self-service portal
as well as manage the life cycle of these applications tax
this is the first of three videos about the application stack accelerator
the other two will focus more on the orchestration and platform specific
details
but for now I'd like to explain what we mean by an application stack
and what it takes to automate one before you can automate the creation of an
entire stack you have to have a very clear and detailed design
a reference architecture in our solution we call this the stack blueprint design
you can think of this blueprint as a template from which in use or to order
whole application stacks
or platforms this accelerator
pack introduces a couple of new user roles and you're seeing the home page
for stack consumer
but probably the most important role is that a the stack
blueprint designer and Here I am logged in as someone playing that role
typically this would be an enterprise architect responsible for defining the
standard platforms
most commonly needed to run the business the first task at hand is to define a
set a target environments
now I've created some already but let's take a look at what the process involves
now since the precise details of your blueprint will vary depending upon the
target hypervisor
and whether using sheffer puppet to perform server configuration
it's easier if you first identify your likely targets for orchestration
so I'm gonna create RV Center in puppet
so here are defined design target that uses the center is the computer good
and puppet as a configuration target I can now use this target by name in any
blueprint I create
so let's take a closer look at creating a blueprint after is to supply the basic
details at my blueprint such as its name in business purpose
and of course the target environment I get to the heart of the matter
specifying the servers in that stack
and now we see why the definition of a design target
is so important because I selected avi Center
puppet combination I'm asked to specify which
the center template should be applied to each server and likewise the operating
system
list is derived automatically from the templates that are available
these are retrieved through what we call the cloud sync process
which will go into more detail on the subsequent videos
after specifying what type of server this is
which is really more informational to the end-user
I'm prompted to select from standardized
machine sizes now
both ec2 an OpenStack have the notional
image sizes are flavors so if I had instead
Pepsi and OpenStack target I'd be prompted with a different list here
let me just do that of course the search to OpenStack
also means that I have to specify different VM template
but in any case the total the CPU and memory requirements for the entire stack
will be due
derived automatically when I complete the blueprint design
so at this point I need to specify the details for all the remaining servers in
the stack
defining the servers in the stacker the computer component
is just one dimensional stack blueprint design equally important
is the definition of the networking infrastructure required to run the stack
in our solution we've defined for standard types that network sounds
into which a server can be placed now obviously there may be more than four
types and zones
but these represent the most common use cases and cloud computing
and you can get a %uh view of what they represent by hovering here
now again picking an option from the drop-down list is easy
that's because the complexity of what I'm doing by picking a value here
is handled by the application as all show in a minute now
if I were creating a three-tiered architecture chances are that I might
want to design for high availability at the web here
by specifying a maximum number here
I'm allowing for that possibility
so one last been a business
regarding the servers because I've define multiple
in this blueprint I need to specify a boot sequence that will apply when the
entire stack it's powered up
or power cycle that any on later point let me just do that
now that I finished with the computer aspect at the blueprint
I can turn my full attention to the networking aspect
a first since I said that the end user can order up to two web servers
I probably need to specify small balancing it's as easy as this
now that my blueprint is really getting interesting here
am let's just take a look at what I it consists of has three servers
one which can be ordered in multiples has a load balancer
with an implicit lower bounds for the next step is defying the firewall rules
that will govern how the tears in the stack communicate with each other
whether they're in the same networks on a different network sounds because I
stated the types of zones into which I want to place the compute nodes
the application can for a few things and recommend some typical firewall rules
for me to use
I can choose to accept those suggestions
or I can just go ahead and specify some
my own
again this process looks easy
and yet building up the specification to a detail that can be orchestrated
is nothing but now we get to the configuration management
dimension and this is who where I can and sheffer puppet rules to each of the
servers
in my stack now because I selected public earlier
it's showing me the public roles that are available
this list roles as well as the individual cookbooks eat roll contains
and the attributes required by each of those comes automatically through our
cloud sync process
as a blueprint designer and probably more interested in constraining these
attributes to standard values
but there may be instances where I truly need input from the end user
up
so I'm able to specify here default values for each
attribute for eat role and for each server in the stack
but also whether the end user who orders from this group print will get prompted
for value
finally I get to the fun part
our solution captures all the state about a blueprint design in XML form to
standard known as Tosca
the task specification provides a very helpful framework for
representing the blueprint design and transforming that design
into something consumable by an orchestrator so when I click this button
you can see that I end up with a list of all the components in my blueprint
along with the key relationships amongst those components
and finally listing requirements this last section is where we see everything
about this blueprint design that has been inferred from what I answered
for example when we go to instantiate the stack
only define networks matching his own types that I specified for the server
placement
will also need to find a resource pool
that will accommodate the total CPU's amram required by all the servers
and because I have more than one server in the stack I'm gonna need a router
because I have one firewall rule within one's own
I'm gonna need a computer firewall and because I have another firewall rule
that crosses arms I'm going to need an edge firewall
all this is captured by the blueprint design
now I just shown a stack using public roles aimed at the center
I mean may need a number variations on this theme
perhaps a similar stack design using chef instead
for one aimed at OpenStack or maybe I need exactly the same stack design
but with smaller servers safer test environment
that's where solution makes it easy for you to copy existing blueprints
and edit them accordingly
in fact we provide a full set of services to manage the life cycle other
step blueprint from creating a test ACK from it
and editing it accordingly to getting a review for america textual governments
team
and when you're finally ready to publish it it's a simple matter I'm using this
link
now with my three-tiered blueprint published
end-users or step consumers can order instances ovett
to my specifications the next video will show what it looks like to order an
instance of this multi-tiered cloud application stack
to learn more about Cisco's cloud accelerators
or to download the application stack accelerator please go to the website
shown on the screen Designing Blueprints for Cloud Applications

Applications in the cloud Designing Blueprints for Cloud Applications By Applications in the cloud Applications in the cloud Published: 2014-10-03T17:50:00-07:00 By Applications in the cloud: Designing Blueprints for Cloud Applications Applications in the cloud Published: 2014-10-03T17:50:00-07:00 Designing Blueprints for Cloud Applications 5 99909 reviewer
Title : Designing Blueprints for Cloud Applications
Description : hello i'm katherine word from Cisco's cloud and Systems Management Technology Group here to present the latest innovation within our...
Rating : 5 CA California Located at
123 Main St , California , 36.778261 -119.4179324 CA . Phone: 206-555-1234 http://california.com What is the latitude and longitude of the Applications in the cloud ? Answer:
http://schema.org/GeoCoordinates