Cloud computing Flashcards
Users expect a cloud infrastructure to support:
The illusion of infinite computing resources available on-demand
The elimination of up-front commitment by users
The ability to pay for use of computing resources on a short-term, as-needed basis
Cloud Components: Clients
Clients are the devices used to interact and manage information on the cloud.
They generally fall into 3 categories
Mobile. Devices such as IPads, iPhone, SmartPhone.
Thin. Clients are computers with no storage, allow the server to do all the work and simply display information.
Thick. Clients are regular computers using browsers to
connect to the Cloud.
Thin clients: Advantages
Lower Hardware Costs
Lower IT Costs
Security
Data Security (in the event of hardware failure)
Less Power consumption
Ease of Repair or Replacement
Less Noise (no fans)
Cloud Components: Datacenter
The datacenter is the collection of servers where the target application is housed.
Its location is varied: a basement, a server farm, a remote location or country.
A growing trend is virtualising servers.
Multiple instances of the same server can be used on a single physical server.
The number of virtual servers depends on the size and speed of the physical server and the types of applications
Cloud Components: Distributed Servers
Servers can be in different geographic locations but this is hidden from the end-user.
This gives the service provider more flexibility in options and security.
Amazon maintain their cloud solution in servers around the world. If one site fails, access is provided through a second site.
If more computing power (hardware) is required, it is not necessary to add more to their server room. Instead, they can be added to any location and made part of the cloud.
Infrastructure
There is no single configuration: a Cloud Computing infrastructure can be deployed in several ways.
This is one of the benefits of the cloud.
If your requirements are large scale and possibly too big for your organisation, use the Cloud Computing services.
If your requirements are tiny and not worth buying the infrastructure, use the Cloud Computing services.
Grid Infrastructure
Grid Computing is often confused with Cloud Computing: they are quite different!
Grid Computing applies the resources of numerous computers in a network to work on a “single problem simultaneously”.
Grid vs Cloud
Grid Computing uses software that divides a problem among thousands of computers, either throughout an organisation’s computers or in the form of a public collaboration.
Cloud Computing is the opposite: it facilitates multiple smaller applications to run at the same time.
Full Virtualisation
Full Virtualisation is a technique where a complete installation of one machine runs on another.
The result is a system where all software running on the server is within the virtual machine.
Full Virtualisation Benefits
Virtualisation is important to Cloud Computing as it provides one method of accessing services on the cloud.
The remote datacenter delivers services in a fully virtualised format.
Full Virtualisation has been successful for a number of reasons:
Sharing a computer system among multiple users
Isolating users from each other and from the control program
Emulating hardware on another machine
Paravirtualisation
Paravirtualisation allows multiple O/S to run on a single hardware device simultaneously by being more efficient at using resources such as processors and memory.
In Full Virtualisation, the entire system is emulated but in Para, its management center works with a customised O/S (modified for a virtual machine).
The different systems are aware of each other and cooperate for increased efficiency.
Para runs better than Full because in Full, all elements must be emulated.
. . . in other words full emulation (virtualisation).
Guest programs must be specifically modified to run in this environment.
Paravirtualisation: The tradeoff
The trade-off is reduced security and flexibility :
Flexibility is reduced as new versions of guest O/S may not work.
The underlying O/S was adjusted to work with the original version of the guest O/S.
Security is compromised as the guest O/S has more control of underlying hardware.
In Full virtualisation, there is effectively a wall between guests and the underlying O/S, and also each other.
Paravirtualisation Benefits
Paravirtualisation works best in specific deployments.
Disaster Recovery. In the event of failure, guest instances can be moved to other hardware until the failure is resolved.
Migration. Moving to a new system is easier and faster as guest instances can be removed from the underlying hardware.
Capacity Management. Because of easier migrations,
capacity management is easier to implement. It is easy to add more processing or hard drive capacity in the paravirtualisation environment.
Full Virtualisation also provides all of the above!
Myth no.1
My Data Center Is Virtualized So I Already Have the Cloud
Virtualisation is a key component of the cloud, but Cloud Computing requires far more than virtualisation.
What defines the Cloud is the ability to deploy and scale infrastructure:
rapidly and programmatically
on-demand
on a pay-as-you-go basis.
Proper Load Balancing and Sharing!
This is difficult if not impossible to achieve using traditional virtualisation alone.
Myth no.2
Cloud Application Performance Is Hampered by Network Latency and I/O Bottlenecks
This is a deployment issue!
Both can be issues, but the degree to which they affect performance depends on:
cloud providers,
network providers,
applications,
cloud deployment architecture.
A carefully architected deployment can often avoid these issues.