What is Cloud? amazonaws com
1
To a large extent, the term Cloud is very ‘cloudy’ to many people (Excuse the bad pun). The understanding of Cloud ranges from it being infrastructure that is not managed by us, infrastructure which at a remote location all the way to Cloud being an utility service. There is a certain amount of truth in what people believe but it is no the whole truth.
We need to clearly understand what are the differences between Cloud and a Data Center. Cloud is run from a Data Center but every Data Center is not a Cloud. Once we understand what Cloud is in clear terms, the difference between these two will become obvious and the confusion will disappear.
2
Before defining what is a Cloud, it will be good if we understand various types of Cloud(s). We can look at Cloud along two axes. One, how is the Cloud deployed (Deployment Model) . Two, what service does the Cloud provide us (Service Model). These two are the popular paradigms of the Cloud and let us understand them since these terms will be heard often. Deployment models include: - Public Cloud - Private Cloud - Hybrid Cloud Service models include: - Infrastructure as a Service (IaaS) - Platform as a Service (PaaS) - Software as a Service (SaaS)
3
Public Cloud: as the name indicates, is available to the general public. Anyone can sign up and use the services of the Cloud Provider. Amazon AWS is the most famous example but it is not the only one. You have others like Microsoft Azure, IBM SoftLayer etc which are Public Clouds. Private Cloud: is owned (or leased) by a company for its exclusive use. Companies like RackSpace help you in setting up your own Private Cloud. Hybrid Cloud: Lot of definitions exist but the one which seems logical is that any cloud made up of two or more clouds is a Hybrid Cloud. Assume you have a private cloud in which you store your data and you want to store your backups and archives in a Public cloud and you connect these two together, then you get a Hybrid Cloud. This is a good deployment -type if you do not want your critical data to lie in any multi-tenancy storage or if are unsure of public cloud’s security. Companies like Microsoft Azure promise that they can help companies build Hybrid clouds easily.
4
Have a look at the picture carefully. Many people find it strange that there can be self hosted and self managed Private Cloud !!! This goes against their standard understanding that if it is Cloud, it must be managed and hosted by ‘someone else’. And how can a Public Cloud be ‘dedicated’? All these queries will be answered once we understand how Cloud is defined.
5
Cloud is also divided based on the Service that it offers. Your Cloud can be public, private or hybrid and each one of them can offer any of these services. While people tend to get fancy with service names, the generally accepted division based on services is as follows: - Infrastructure as a Service (IaaS) - Platform as a Service (PaaS) - Software as a Service (SaaS) The diagram in the slide clearly shows us what each service is all about. It all depends on level of control the user has. In case of IaaS, the user control everything from the Operating System upwards. The Cloud provide provides the Compute, Storage and the Networking required and the user chooses the Operating System, the tools, the middleware and everything else. In case of PaaS, the user is in control of what application she wants to deploy and the data. The provider will provide the platform which would include the middleware, compilation tools, IDEs etc. The user develops and deploys an application based on the platform chosen. In case of SaaS, everything is controlled by the provider The user uses the software which is available for her purpose.
6
Amazon AWS’s EC2 and S3 are good examples of IaaS, Google App Engine is a good example of PaaS and Office 365 and Salesforce.com are good examples of SaaS
6
In order to clear the confusion about what is Cloud, we should look at the definition provided by National Institute of Standards and Technology (NIST). The five key characteristics enumerated here will provide for an unambiguous understanding of the Cloud. The five characteristics are: On-Demand Self Service Broad Network Access Resource Pooling Rapid Elasticity Measured Service
7
One of the key characteristics of the Cloud is On Demand Self Service. In most enterprises, procuring server or storage resource is a long process. It starts by the user raising a Capital Asset Request and goes through an approval chain. After approval, the ordering process starts and when the resource would arrive depends a lot on your luck. In case of Cloud, the user can provision the required compute or storage whenever required and the resource would be available almost instantly. It does not require a department to procure the resource and a system/storage administrator to provision the resource. The procurement and provisioning are both done by the user thus saving on time and administrative cost. The other important aspect to be considered is that these resources can be released by the user at any point in time.
8
The Cloud must be accessible via the network and in many cases from the internet. Generally all Cloud Provider provide REST APIs to access their services. The REST APIs generally work over HTTP and hence the Cloud Providers must support HTTP and HTTPS. In the modern world, we connect to the net not just via Desktop or Laptops. We use multiple devices like Mobile Phones, Tablets etc to connect to the net and going forward these devices could multiply considering that everyone is talking about Internet of Things (IOT) So Cloud Providers must ensure that users can connect to their services via multiple devices by providing the necessary apps.
9
This is what ensures that the user gets a good deal from the Cloud Provider. The Cloud Provider would pool the resources and ensure all resources are used optimally and there is no wastage. This is not possible in pure Enterprise Data Center type of setup where different divisions own different resources and the resource boundaries are strictly enforced.
Resource Pooling automatically implies multi tenancy and for corporates with strict security requirements this could be an issue. The resources in the pool are allocated dynamically to users. By pooling resources and ensuring optimal usage of resources, the Cloud Provider is able to reduce the overall cost and the benefit is passed to the user.
10
This is one more of the key value proposition of the Cloud, a feature which is used to sell the Cloud. The ability to rapidly grow or rapidly shrink is a boon to many users. This means that you have not locked in any Capital Asset which you are not using now. At the same time you are not delaying a project since you have not bought the Capital Asset yet or you have underestimated resource needs. You can get what you what and grow as and when you need. The Cloud assures you this elasticity. The Cloud allows you to increase your resources at a click of a button and also allows you to release the resources at the click of a button. Scaling up and scaling down are now a breeze. (Ofcourse companies must exercise some sort of control over this since you do not want Shadow IT happening in your organization)
11
This is what people understand as Utility Service. You are billed for what you use. It is similar to how you are billed for Electricity or Water. You don’t need to commit before hand on any resource. For example, take the case of Storage. You start storing your files / objects and the Storage keeps scaling as more objects are being stored. You will be billed on the exact amount of space your objects occupy. If you delete some objects, the billing would reduce. If you add more objects you billing would increase. The granularity of billing will depend on the resource used and the service provider.
12
This slide explains how a Data Center becomes a Cloud by adding the attributes we spoke about till now. If you can setup your data center to have all the 5 attributes we discussed, then your data center can be called a Cloud. So it doesn’t matter who owns or operates the Data Center. What matter is, does it have: - Automatic Self Provisioning - Broad Network Access - Resource Pooling - Elasticity - Usage based billing If these are present, then it is a Cloud. It can be in your own premises, it can be in your partners premises or it can be half way across the globe. It may be managed by your company or b a third party. It doesn’t matter. If you have the 5 characteristics mentioned, then your Data Center is a Cloud Data Center. Else it is just a Data Center, maybe managed by a third party but still only a Data Center and not a Private Cloud.
13
The basic Cloud infrastructure consists of hypervisors (Xen in case of Amazon) which can launch Virtual Machines of different OS like Linux or Windows, Virtualized Storage and a Virtual Network. These form the core infrastructure. On top of it would the management framework which allows the users to interact with the Cloud. The framework would contain APIs which connect and manage the infrastructure via APIs. The framework will also have the console and command line interfaces which are required by the user to manage the Cloud.
14
15
To a large extent, the term Cloud is very ‘cloudy’ to many people (Excuse the bad pun). The understanding of Cloud ranges from it being infrastructure that is not managed by us, infrastructure which at a remote location all the way to Cloud being an utility service. There is a certain amount of truth in what people believe but it is no the whole truth.
We need to clearly understand what are the differences between Cloud and a Data Center. Cloud is run from a Data Center but every Data Center is not a Cloud. Once we understand what Cloud is in clear terms, the difference between these two will become obvious and the confusion will disappear.
2
Before defining what is a Cloud, it will be good if we understand various types of Cloud(s). We can look at Cloud along two axes. One, how is the Cloud deployed (Deployment Model) . Two, what service does the Cloud provide us (Service Model). These two are the popular paradigms of the Cloud and let us understand them since these terms will be heard often. Deployment models include: - Public Cloud - Private Cloud - Hybrid Cloud Service models include: - Infrastructure as a Service (IaaS) - Platform as a Service (PaaS) - Software as a Service (SaaS)
3
Public Cloud: as the name indicates, is available to the general public. Anyone can sign up and use the services of the Cloud Provider. Amazon AWS is the most famous example but it is not the only one. You have others like Microsoft Azure, IBM SoftLayer etc which are Public Clouds. Private Cloud: is owned (or leased) by a company for its exclusive use. Companies like RackSpace help you in setting up your own Private Cloud. Hybrid Cloud: Lot of definitions exist but the one which seems logical is that any cloud made up of two or more clouds is a Hybrid Cloud. Assume you have a private cloud in which you store your data and you want to store your backups and archives in a Public cloud and you connect these two together, then you get a Hybrid Cloud. This is a good deployment -type if you do not want your critical data to lie in any multi-tenancy storage or if are unsure of public cloud’s security. Companies like Microsoft Azure promise that they can help companies build Hybrid clouds easily.
4
Have a look at the picture carefully. Many people find it strange that there can be self hosted and self managed Private Cloud !!! This goes against their standard understanding that if it is Cloud, it must be managed and hosted by ‘someone else’. And how can a Public Cloud be ‘dedicated’? All these queries will be answered once we understand how Cloud is defined.
5
Cloud is also divided based on the Service that it offers. Your Cloud can be public, private or hybrid and each one of them can offer any of these services. While people tend to get fancy with service names, the generally accepted division based on services is as follows: - Infrastructure as a Service (IaaS) - Platform as a Service (PaaS) - Software as a Service (SaaS) The diagram in the slide clearly shows us what each service is all about. It all depends on level of control the user has. In case of IaaS, the user control everything from the Operating System upwards. The Cloud provide provides the Compute, Storage and the Networking required and the user chooses the Operating System, the tools, the middleware and everything else. In case of PaaS, the user is in control of what application she wants to deploy and the data. The provider will provide the platform which would include the middleware, compilation tools, IDEs etc. The user develops and deploys an application based on the platform chosen. In case of SaaS, everything is controlled by the provider The user uses the software which is available for her purpose.
6
Amazon AWS’s EC2 and S3 are good examples of IaaS, Google App Engine is a good example of PaaS and Office 365 and Salesforce.com are good examples of SaaS
6
In order to clear the confusion about what is Cloud, we should look at the definition provided by National Institute of Standards and Technology (NIST). The five key characteristics enumerated here will provide for an unambiguous understanding of the Cloud. The five characteristics are: On-Demand Self Service Broad Network Access Resource Pooling Rapid Elasticity Measured Service
7
One of the key characteristics of the Cloud is On Demand Self Service. In most enterprises, procuring server or storage resource is a long process. It starts by the user raising a Capital Asset Request and goes through an approval chain. After approval, the ordering process starts and when the resource would arrive depends a lot on your luck. In case of Cloud, the user can provision the required compute or storage whenever required and the resource would be available almost instantly. It does not require a department to procure the resource and a system/storage administrator to provision the resource. The procurement and provisioning are both done by the user thus saving on time and administrative cost. The other important aspect to be considered is that these resources can be released by the user at any point in time.
8
The Cloud must be accessible via the network and in many cases from the internet. Generally all Cloud Provider provide REST APIs to access their services. The REST APIs generally work over HTTP and hence the Cloud Providers must support HTTP and HTTPS. In the modern world, we connect to the net not just via Desktop or Laptops. We use multiple devices like Mobile Phones, Tablets etc to connect to the net and going forward these devices could multiply considering that everyone is talking about Internet of Things (IOT) So Cloud Providers must ensure that users can connect to their services via multiple devices by providing the necessary apps.
9
This is what ensures that the user gets a good deal from the Cloud Provider. The Cloud Provider would pool the resources and ensure all resources are used optimally and there is no wastage. This is not possible in pure Enterprise Data Center type of setup where different divisions own different resources and the resource boundaries are strictly enforced.
Resource Pooling automatically implies multi tenancy and for corporates with strict security requirements this could be an issue. The resources in the pool are allocated dynamically to users. By pooling resources and ensuring optimal usage of resources, the Cloud Provider is able to reduce the overall cost and the benefit is passed to the user.
10
This is one more of the key value proposition of the Cloud, a feature which is used to sell the Cloud. The ability to rapidly grow or rapidly shrink is a boon to many users. This means that you have not locked in any Capital Asset which you are not using now. At the same time you are not delaying a project since you have not bought the Capital Asset yet or you have underestimated resource needs. You can get what you what and grow as and when you need. The Cloud assures you this elasticity. The Cloud allows you to increase your resources at a click of a button and also allows you to release the resources at the click of a button. Scaling up and scaling down are now a breeze. (Ofcourse companies must exercise some sort of control over this since you do not want Shadow IT happening in your organization)
11
This is what people understand as Utility Service. You are billed for what you use. It is similar to how you are billed for Electricity or Water. You don’t need to commit before hand on any resource. For example, take the case of Storage. You start storing your files / objects and the Storage keeps scaling as more objects are being stored. You will be billed on the exact amount of space your objects occupy. If you delete some objects, the billing would reduce. If you add more objects you billing would increase. The granularity of billing will depend on the resource used and the service provider.
12
This slide explains how a Data Center becomes a Cloud by adding the attributes we spoke about till now. If you can setup your data center to have all the 5 attributes we discussed, then your data center can be called a Cloud. So it doesn’t matter who owns or operates the Data Center. What matter is, does it have: - Automatic Self Provisioning - Broad Network Access - Resource Pooling - Elasticity - Usage based billing If these are present, then it is a Cloud. It can be in your own premises, it can be in your partners premises or it can be half way across the globe. It may be managed by your company or b a third party. It doesn’t matter. If you have the 5 characteristics mentioned, then your Data Center is a Cloud Data Center. Else it is just a Data Center, maybe managed by a third party but still only a Data Center and not a Private Cloud.
13
The basic Cloud infrastructure consists of hypervisors (Xen in case of Amazon) which can launch Virtual Machines of different OS like Linux or Windows, Virtualized Storage and a Virtual Network. These form the core infrastructure. On top of it would the management framework which allows the users to interact with the Cloud. The framework would contain APIs which connect and manage the infrastructure via APIs. The framework will also have the console and command line interfaces which are required by the user to manage the Cloud.
14
15
