1. Discuss the benefits and constraints of different networkingsystems types and topologies.

One way to categorize the different types of computer network designs is by their scope or scale. For
historical reasons, the networking industry refers to nearly every type of design as some kind of area
network. Common examples of area network types are:
LAN - Local Area Network
WLAN - Wireless Local Area Network
WAN - Wide Area Network
MAN - Metropolitan Area Network
SAN - Storage Area Network, System Area Network, Server Area Network, or sometimes Small Area
Network
CAN - Campus Area Network, Controller Area Network, or sometimes Cluster Area Network
PAN - Personal Area Network
DAN - Desk Area Network
LAN and WAN were the original categories of area networks, while the others have gradually emerged
over many years of technology evolution.
Note that these network types are a separate concept from network topologies such as bus, ring and
star.

See also - Introduction to Network Topologies
LAN - Local Area Network

A LAN connects network devices over a relatively short distance. A networked office building, school, or
home usually contains a single LAN, though sometimes one building will contain a few small LANs
(perhaps one per room), and occasionally a LAN will span a group of nearby buildings. In TCP/IP
networking, a LAN is often but not always implemented as a single IP subnet.
In addition to operating in a limited space, LANs are also typically owned, controlled, and managed by a
single person or organization. They also tend to use certain connectivity technologies, primarily Ethernet
and Token Ring.

WAN - Wide Area Network

As the term implies, a WAN spans a large physical distance. The Internet is the largest WAN, spanning
the Earth.
A WAN is a geographically-dispersed collection of LANs. A network device called a router connects LANs
to a WAN. In IP networking, the router maintains both a LAN address and a WAN address.

A WAN differs from a LAN in several important ways. Most WANs (like the Internet) are not owned by
any one organization but rather exist under collective or distributed ownership and management. WANs
tend to use technology like ATM, Frame Relay and X.25 for connectivity over the longer distances.

LAN, WAN and Home Networking

Residences typically employ one LAN and connect to the Internet WAN via an Internet Service Provider
(ISP) using a broadband modem. The ISP provides a WAN IP address to the modem, and all of the
computers on the home network use LAN (so-called private) IP addresses. All computers on the home
LAN can communicate directly with each other but must go through a central gateway, typically a
broadband router, to reach the ISP.
Other Types of Area Networks

While LAN and WAN are by far the most popular network types mentioned, you may also commonly see
references to these others:
Wireless Local Area Network - a LAN based on WiFi wireless network technology
Metropolitan Area Network - a network spanning a physical area larger than a LAN but smaller than a
WAN, such as a city. A MAN is typically owned an operated by a single entity such as a government body
or large corporation.
Campus Area Network - a network spanning multiple LANs but smaller than a MAN, such as on a
university or local business campus.
Storage Area Network - connects servers to data storage devices through a technology like Fibre
Channel.
System Area Network - links high-performance computers with high-speed connections in a cluster
configuration. Also known as Cluster Area Network.

Geographic scale[edit]
A network can be characterized by its physical capacity or its organizational purpose. Use of the network,
including user authorization and access rights, differ accordingly.
Personal area network
A personal area network (PAN) is a computer network used for communication among computer and
different information technological devices close to one person. Some examples of devices that are used
in a PAN are personal computers, printers, fax machines, telephones, PDAs, scanners, and even video
game consoles. A PAN may include wired and wireless devices. The reach of a PAN typically extends to
10 meters.
[15]
A wired PAN is usually constructed with USB and FireWire connections while technologies
such as Bluetooth and infrared communication typically form a wireless PAN.
Local area network
A local area network (LAN) is a network that connects computers and devices in a limited geographical
area such as a home, school, office building, or closely positioned group of buildings. Each computer or
device on the network is a node. Wired LANs are most likely based on Ethernet technology. Newer
standards such as ITU-T G.hn also provide a way to create a wired LAN using existing wiring, such as
coaxial cables, telephone lines, and power lines.
[16]

A LAN is depicted in the accompanying diagram. All interconnected devices use the network layer (layer
3) to handle multiple subnets(represented by different colors). Those inside the library have 10/100 Mbit/s
Ethernet connections to the user device and a Gigabit Ethernet connection to the central router. They
could be called Layer 3 switches, because they only have Ethernet interfaces and support the Internet
Protocol. It might be more correct to call them access routers, where the router at the top is a distribution
router that connects to the Internet and to the academic networks' customer access routers.
The defining characteristics of a LAN, in contrast to a wide area network (WAN), include higher data
transfer rates, limited geographic range, and lack of reliance on leased lines to provide connectivity.
Current Ethernet or other IEEE 802.3 LAN technologies operate at data transfer rates up to 10 Gbit/s.
The IEEE investigates the standardization of 40 and 100 Gbit/s rates.
[17]
A LAN can be connected to a
WAN using a router.
Home area network
A home area network (HAN) is a residential LAN which is used for communication between digital devices
typically deployed in the home, usually a small number of personal computers and accessories, such as
printers and mobile computing devices. An important function is the sharing of Internet access, often a
broadband service through a cable TV or digital subscriber line (DSL) provider.
Storage area network
A storage area network (SAN) is a dedicated network that provides access to consolidated, block level
data storage. SANs are primarily used to make storage devices, such as disk arrays, tape libraries, and
optical jukeboxes, accessible to servers so that the devices appear like locally attached devices to the
operating system. A SAN typically has its own network of storage devices that are generally not
accessible through the local area network by other devices. The cost and complexity of SANs dropped in
the early 2000s to levels allowing wider adoption across both enterprise and small to medium sized
business environments.
Campus area network
A campus area network (CAN) is made up of an interconnection of LANs within a limited geographical
area. The networking equipment (switches, routers) and transmission media (optical fiber, copper
plant, Cat5 cabling, etc.) are almost entirely owned by the campus tenant / owner (an enterprise,
university, government, etc.).
For example, a university campus network is likely to link a variety of campus buildings to connect
academic colleges or departments, the library, and student residence halls.
Backbone network
A backbone network is part of a computer network infrastructure that provides a path for the exchange of
information between different LANs or sub-networks. A backbone can tie together diverse networks within
the same building, across different buildings, or over a wide area.
For example, a large company might implement a backbone network to connect departments that are
located around the world. The equipment that ties together the departmental networks constitutes the
network backbone. When designing a network backbone,network performance and network
congestion are critical factors to take into account. Normally, the backbone network's capacity is greater
than that of the individual networks connected to it.
Another example of a backbone network is the Internet backbone, which is the set of wide area
networks (WANs) and core routers that tie together all networks connected to the Internet.
Metropolitan area network
A Metropolitan area network (MAN) is a large computer network that usually spans a city or a large
campus.
Wide area network
A wide area network (WAN) is a computer network that covers a large geographic area such as a city,
country, or spans even intercontinental distances. A WAN uses a communications channel that combines
many types of media such as telephone lines, cables, and air waves. A WAN often makes use of
transmission facilities provided by common carriers, such as telephone companies. WAN technologies
generally function at the lower three layers of the OSI reference model: the physical layer, the data link
layer, and the network layer.
Enterprise private network
An enterprise private network is a network built by a single organization to interconnect its office locations
(e.g., production sites, head offices, remote offices, shops) in order to share computer resources.
Virtual private network
A virtual private network (VPN) is an overlay network in which some of the links between nodes are
carried by open connections or virtual circuits in some larger network (e.g., the Internet) instead of by
physical wires. The data link layer protocols of the virtual network are said to be tunneled through the
larger network when this is the case. One common application is secure communications through the
public Internet, but a VPN need not have explicit security features, such as authentication or content
encryption. VPNs, for example, can be used to separate the traffic of different user communities over an
underlying network with strong security features.
VPN may have best-effort performance, or may have a defined service level agreement (SLA) between
the VPN customer and the VPN service provider. Generally, a VPN has a topology more complex than
point-to-point.
Global area network
A global area network (GAN) is a network used for supporting mobile across an arbitrary number of
wireless LANs, satellite coverage areas, etc. The key challenge in mobile communications is handing off
user communications from one local coverage area to the next. In IEEE Project 802, this involves a
succession of terrestrial wireless LANs.
[18]

Organizational scope[edit]
Networks are typically managed by the organizations that own them. Private enterprise networks may use
a combination of intranets and extranets. They may also provide network access to the Internet, which
has no single owner and permits virtually unlimited global connectivity.
Intranets[edit]
An intranet is a set of networks that are under the control of a single administrative entity. The intranet
uses the IP protocol and IP-based tools such as web browsers and file transfer applications. The
administrative entity limits use of the intranet to its authorized users. Most commonly, an intranet is the
internal LAN of an organization. A large intranet will typically have at least one web server to provide
users with organizational information.
Extranet[edit]
An extranet is a network that is also under the administrative control of a single organization, but supports
a limited connection to a specific external network. For example, an organization may provide access to
some aspects of its intranet to share data with its business partners or customers. These other entities
are not necessarily trusted from a security standpoint. Network connection to an extranet is often, but not
always, implemented via WAN technology.
Internetwork[edit]
An internetwork is the connection of multiple computer networks via a common routing technology using
routers.
Internet[edit]


Partial map of the Internet based on the January 15, 2005 data found on opte.org. Each line is drawn between two nodes,
representing two IP addresses. The length of the lines are indicative of the delay between those two nodes. This graph
represents less than 30% of the Class Cnetworks reachable.
The Internet is the largest example of an internetwork. It is a global system of interconnected
governmental, academic, corporate, public, and private computer networks. It is based on the networking
technologies of the Internet Protocol Suite. It is the successor of the Advanced Research Projects Agency
Network (ARPANET) developed by DARPA of theUnited States Department of Defense. The Internet is
also the communications backbone underlying the World Wide Web (WWW).
Participants in the Internet use a diverse array of methods of several hundred documented, and often
standardized, protocols compatible with the Internet Protocol Suite and an addressing system (IP
addresses) administered by the Internet Assigned Numbers Authority and address registries. Service
providers and large enterprises exchange information about the reachability of their address spaces
through the Border Gateway Protocol (BGP), forming a redundant worldwide mesh of transmission paths.
Darknet[edit]
A Darknet is an overlay network, typically running on the internet, that is only accessible through
specialized software. A darknet is an anonymizing network where connections are made only between
trusted peers sometimes called "friends" (F2F)
[19]
using non-standard protocols and ports.
Darknets are distinct from other distributed peer-to-peer networks as sharing is anonymous (that is, IP
addresses are not publicly shared), and therefore users can communicate with little fear of governmental
or corporate interference.
[20]


What is Bus topology?
Bus Topology is the simplest of network topologies. In this type of topology, all the nodes (computers
as well as servers) are connected to the single cable (called bus), by the help of interface
connectors. This central cable is the backbone of the network and is known as Bus (thus the name).
Every workstation communicates with the other device through this Bus.

A signal from the source is broadcasted and it travels to all workstations connected to bus cable.
Although the message is broadcasted but only the intended recipient, whose MAC address or IP
address matches, accepts it. If the MAC /IP address of machine doesnt match with the intended
address, machine discards the signal.

A terminator is added at ends of the central cable, to prevent bouncing of signals. A barrel connector
can be used to extend it. Below I have given a basic diagram of a bus topology and then have
discussed advantages and disadvantages of Bus Network Topology


Bus topology diagram


Advantages (benefits) of Linear Bus Topology

1) It is easy to set-up and extend bus network.
2) Cable length required for this topology is the least compared to other networks.
3) Bus topology costs very less.
4) Linear Bus network is mostly used in small networks. Good for LAN.


Disadvantages (Drawbacks) of Linear Bus Topology

1) There is a limit on central cable length and number of nodes that can be connected.
2) Dependency on central cable in this topology has its disadvantages.If the main cable (i.e. bus )
encounters some problem, whole network breaks down.
3) Proper termination is required to dump signals. Use of terminators is must.
4) It is difficult to detect and troubleshoot fault at individual station.
5) Maintenance costs can get higher with time.
6) Efficiency of Bus network reduces, as the number of devices connected to it increases.
7) It is not suitable for networks with heavy traffic.
8) Security is very low because all the computers receive the sent signal from the source.

What is Star topology?
In Star topology, all the components of network are connected to the central device called hub
which may be a hub, a router or a switch. Unlike Bus topology (discussed earlier), where nodes were
connected to central cable, here all the workstations are connected to central device with a point-to-
point connection. So it can be said that every computer is indirectly connected to every other node
by the help of hub.

All the data on the star topology passes through the central device before reaching the intended
destination. Hub acts as a junction to connect different nodes present in Star Network, and at the
same time it manages and controls whole of the network. Depending on which central device is
used, hub can act as repeater or signal booster. Central device can also communicate with other
hubs of different network. Unshielded Twisted Pair (UTP) Ethernet cable is used to connect
workstations to central node.


Star Topology Diagram


Advantages of Star Topology
1) As compared to Bus topology it gives far much better performance, signals dont necessarily get
transmitted to all the workstations. A sent signal reaches the intended destination after passing
through no more than 3-4 devices and 2-3 links. Performance of the network is dependent on the
capacity of central hub.
2) Easy to connect new nodes or devices. In star topology new nodes can be added easily without
affecting rest of the network. Similarly components can also be removed easily.
3) Centralized management. It helps in monitoring the network.
4) Failure of one node or link doesnt affect the rest of network. At the same time its easy to detect
the failure and troubleshoot it.


Disadvantages of Star Topology
1) Too much dependency on central device has its own drawbacks. If it fails whole network
goes down.
2) The use of hub, a router or a switch as central device increases the overall cost of the
network.
3) Performance and as well number of nodes which can be added in such topology is
depended on capacity of central device.

What is Ring Topology?
In Ring Topology, all the nodes are connected to each-other in such a way that they make a closed
loop. Each workstation is connected to two other components on either side, and it communicates
with these two adjacent neighbors. Data travels around the network, in one direction. Sending and
receiving of data takes place by the help of TOKEN.

Token Passing (in brief) : Token contains a piece of information which along with data is sent by the
source computer. This token then passes to next node, which checks if the signal is intended to it. If
yes, it receives it and passes the empty to into the network, otherwise passes token along with the
data to next node. This process continues until the signal reaches its intended destination.
The nodes with token are the ones only allowed to send data. Other nodes have to wait for an empty
token to reach them. This network is usually found in offices, schools and small buildings.


Ring Topology & token


Advantages of Ring Topology

1) This type of network topology is very organized. Each node gets to send the data when it
receives an empty token. This helps to reduces chances of collision. Also in ring topology all the
traffic flows in only one direction at very high speed.
2) Even when the load on the network increases, its performance is better than that of Bus
topology.
3) There is no need for network server to control the connectivity between workstations.
4) Additional components do not affect the performance of network.
5) Each computer has equal access to resources.


Disadvantages of Ring Topology

1) Each packet of data must pass through all the computers between source and
destination. This makes it slower than Star topology.
2) If one workstation or port goes down, the entire network gets affected.
3) Network is highly dependent on the wire which connects different components.
4) MAUs and network cards are expensive as compared to Ethernet cards and hubs.

What is Mesh Topology?
In a mesh network topology, each of the network node, computer and other devices, are
interconnected with one another. Every node not only sends its own signals but also relays data
from other nodes. In fact a true mesh topology is the one where every node is connected to every
other node in the network. This type of topology is very expensive as there are many redundant
connections, thus it is not mostly used in computer networks. It is commonly used in wireless
networks. Flooding or routing technique is used in mesh topology.
Types of Mesh Network topologies:-


1)Full Mesh Topology:-

Mesh Topology Diagram
In this, like a true mesh, each component is connected to every other component. Even after
considering the redundancy factor and cost of this network, its main advantage is that the network
traffic can be redirected to other nodes if one of the nodes goes down. Full mesh topology is used
only for backbone networks.


2) Partial Mesh Topology:-
This is far more practical as compared to full mesh topology. Here, some of the systems are
connected in similar fashion as in mesh topology while rests of the systems are only connected to 1
or 2 devices. It can be said that in partial mesh, the workstations are indirectly connected to other
devices. This one is less costly and also reduces redundancy.


Advantages of Mesh topology

1) Data can be transmitted from different devices simultaneously. This topology can withstand high
traffic.
2) Even if one of the components fails there is always an alternative present. So data transfer
doesnt get affected.
3) Expansion and modification in topology can be done without disrupting other nodes.


Disadvantages of Mesh topology

1) There are high chances of redundancy in many of the network connections.
2) Overall cost of this network is way too high as compared to other network topologies.
3) Set-up and maintenance of this topology is very difficult. Even administration of the
network is tough.

What is Tree Topology?

Tree Topology integrates the characteristics of Star and Bus Topology. Earlier we saw how in
Physical Star network Topology, computers (nodes) are connected by each other through central
hub. And we also saw in Bus Topology, work station devices are connected by the common cable
called Bus. After understanding these two network configurations, we can understand tree topology
better. In Tree Topology, the number of Star networks are connected using Bus. This main cable
seems like a main stem of a tree, and other star networks as the branches. It is also
called Expanded Star Topology. Ethernet protocol is commonly used in this type of topology. The
diagram below will make it clear.


Tree Topology

Lets discuss the advantages and disadvantages of Tree Topology now.

Advantages of Tree Topology
1. It is an extension of Star and bus Topologies, so in networks where these topologies can't be
implemented individually for reasons related to scalability, tree topology is the best alternative.
2. Expansion of Network is possible and easy.
3. Here, we divide the whole network into segments (star networks), which can be easily managed
and maintained.
4. Error detection and correction is easy.
5. Each segment is provided with dedicated point-to-point wiring to the central hub.
6. If one segment is damaged, other segments are not affected.


Disadvantages of Tree Topology
1. Because of its basic structure, tree topology, relies heavily on the main bus cable, if it
breaks whole network is crippled.
2. As more and more nodes and segments are added, the maintenance becomes difficult.
3. Scalability of the network depends on the type of cable used.

What is Hybrid Topology ?

Before starting about Hybrid topology, we saw that a network topology is a connection of various
links and nodes, communicating with each other for transfer of data. We also saw various
advantages and disadvantages of Star, Bus, Ring, Mesh and Tree topologies.

Now lets discuss what Hybrid Network topology is and why it finds its application in Wide Area
Networks. Hybrid, as the name suggests, is mixture of two different things. Similarly in this type of
topology we integrate two or more different topologies to form a resultant topology which has good
points(as well as weaknesses) of all the constituent basic topologies rather than having
characteristics of one specific topology. This combination of topologies is done according to the
requirements of the organization.

For example, if there exists a ring topology in one office department while a bus topology in another
department, connecting these two will result in Hybrid topology. Remember connecting two similar
topologies cannot be termed as Hybrid topology. Star-Ring and Star-Bus networks are most
common examples of hybrid network.

Let's see the benefits and drawbacks of this networking architecture


Hybrid Network Topology Image


Advantages of Hybrid Network Topology

1) Reliable : Unlike other networks, fault detection and troubleshooting is easy in this type of
topology. The part in which fault is detected can be isolated from the rest of network and required
corrective measures can be taken, WITHOUT affecting the functioning of rest of the network.
2) Scalable: Its easy to increase the size of network by adding new components, without disturbing
existing architecture.
3) Flexible: Hybrid Network can be designed according to the requirements of the organization and
by optimizing the available resources. Special care can be given to nodes where traffic is high as
well as where chances of fault are high.
4) Effective: Hybrid topology is the combination of two or more topologies, so we can design it in
such a way that strengths of constituent topologies are maximized while there weaknesses are
neutralized. For example we saw Ring Topology has good data reliability (achieved by use of
tokens) and Star topology has high tolerance capability (as each node is not directly connected to
other but through central device), so these two can be used effectively in hybrid star-ring topology.


Disadvantages of Hybrid Topology

1) Complexity of Design: One of the biggest drawback of hybrid topology is its design. Its not easy
to design this type of architecture and its a tough job for designers. Configuration and installation
process needs to be very efficient.
2) Costly Hub: The hubs used to connect two distinct networks, are very expensive. These hubs
are different from usual hubs as they need to be intelligent enough to work with different
architectures and should be function even if a part of network is down.
3) Costly Infrastructure: As hybrid architectures are usually larger in scale, they require a lot of
cables, cooling systems, sophisticate network devices, etc.





What is Cloud Computing? Its advantages &
disadvantages
Since the advent of the Internet, businesses have been introduced to a whole new avenue in their
day-to-day interactions. With the instant ability to access files and information in real time and
anywhere in the world, people across the globe are implementing cloud computing into their
organisations.
While cloud solutions have been the answer for many, some are still hesitant to utilise this new tool.
While there are advantages and disadvantages associated with most business models, it is largely
up to the business owner or IT Department and the nature of the companys daily transactions that
dictate whether this system is right for their industry.
The Advantages
Mobility
One of the main advantages of working in the cloud is that it allows users the mobility necessary in
this day and age of global marketing. For example, a busy executive on a business trip in Japan may
need to know what is going on at the company headquarters in Australia. Rather than having to
make phone calls and rely on the reports of employees, they can simply check updated statistics
online. The Internet is, for the most part, everywhere. Therefore, cloud computing allows the mobility
necessary for success.
Versatile Compatibility
It is an ongoing debate: which is better, the Mac or PC? Despite which side of the fence you stand
on this argument, it makes no difference when it comes to implementing cloud solutions into a
business model. Users are often surprised to find that the various cloud apps available are
accessible on both platforms.
Only Pay for What You Need
Unlike many computing programs where the package comes with unnecessary applications, the
cloud allows users to literally get what they pay for. This scalability allows for you to simply purchase
the applications and data storage you really need.
Individuality
One of the most convenient aspects of working in the cloud is that it is compatible with aspects
specific to the company. For example, cloud IT services can be scaled to meet changing system
demands within a single company.
The Disadvantages
While the cloud benefits are numerous, this method of computation is not for all businesses. There
are certain disadvantages that could persuade you that this system is not for your company, and it
takes careful consideration and professional advice to determine if this is the case in any specific
circumstance.
Less Control
Utilising the public cloud in business does have an obvious downside. By using this technology you
risk losing a level of control over your company. While many IT managers are experimenting with
various ways of implementing an in-house cloud system that runs on delivered metered services,
this is not always the most lucrative business move.
Not Always Enough Room
Many have been disappointed with cloud technology, because they find that once they have
instituted a cloud system within their business, they run out of storage space. While it is possible to
update the system, it can be a painstaking process.
Security and Confidentiality
Since technology has started to expand in the exponential ways we are seeing in this day and age,
cyber-crime has become a concerning issue. Cloud computing does pose the risk of increased
security threats. While most companies have an up-to-date virus database, this does not make the
files and information stored in the cloud immune to hackers.
The extent of these advantages and disadvantages to cloud computing will vary from business to
business, so it will be important for any business to weigh these up when considering their move into
cloud computing. If you need more information or guidance, give us a call, we are happy to help
provide you with the appropriate solution and determine the best IT strategy for you and your
business moving forward.
Cloud Computing Advantages
The advantages and benefits of cloud computing are well documented low to no upfront infrastructure
investments, just in time deployment, and a more efficient resource utilization model are all benefits of
the cloud. Its these very drivers which are creating a significant demand for cloud based services.Major
advantages of cloud computing include:
Major cloud computing advantages include:
Less maintenance: Hardware, applications and bandwidth are managed by the provider.
Continuous availability: Public cloud services are available wherever you are located.
Scalability: Pay only for the applications and data storage you need.
Elasticity: Private clouds can be scaled to meet your changing IT system demands
Expert service: Expedients cloud computing services are continuously monitored and maintained by our
onsite staff of expert data center technicians.
1.2 Advantages of Clusters
Particularly in the area of Internet service deployment, clusters
provide three primary benefits over single larger machines, such as
SMPs: incremental scalability, high availability, and the cost/performance
and maintenance benefits of commodity PCs. We elaborate
on each of these in turn.
Scalability: Clusters are well suited to Internet service workloads,
which are highly parallel (many independent simultaneous
users) and for which the grain size typically corresponds to at most
a few CPU-seconds on a commodity PC. For these workloads,
large clusters can dwarf the power of the largest machines. For
example, Inktomis HotBot cluster contains 60 nodes with 120 processors,
30 GB of physical memory, and hundreds of commodity
disks. Wal-Mart uses a cluster from TeraData with 768 processors
and 16 terabytes of online storage.
Furthermore, the ability to grow clusters incrementally over
time is a tremendous advantage in areas such as Internet service
deployment, where capacity planning depends on a large number
of unknown variables. Incremental scalability replaces capacity
planning with relatively fluid reactionary scaling. Clusters correspondingly
eliminate the forklift upgrade, in which you must
throw out the current machine (and related investments) and
replace it via forklift with an even larger one.
High Availability: Clusters have natural redundancy due to the
independence of the nodes: Each node has its own busses, power
supply, disks, etc., so it is merely a matter of software to mask
(possibly multiple simultaneous) transient faults. A natural extension
of this capability is to temporarily disable a subset of nodes
and then upgrade them in place (hot upgrade). Such capabilities
are essential for network services, whose users have come to
expect 24-hour uptime despite the inevitable reality of hardware
and software faults due to rapid system evolution.
Commodity Building Blocks: The final set of advantages of
clustering follows from the use of commodity building blocks
rather than high-end, low-volume machines. The obvious advantage
is cost/performance, since memory, disks, and nodes can all
track the leading edge; for example, we changed the building block
every time we grew the HotBot cluster, each time picking the reliable
high volume previous-generation commodity units, helping to
ensure stability and robustness. Furthermore, since many commodity
vendors compete on service (particularly for PC hardware), it is
easy to get high-quality configured nodes in 48 hours or less. In
contrast, large SMPs typically have a lead time of 45 days, are
more cumbersome to purchase, install, and upgrade, and are supported
by a single vendor, so it is much harder to get help when difficulties
arise. Once again, it is a simple matter of software to tie
together a collection of possibly heterogeneous commodity building
blocks.
To summarize, clusters have significant advantages in scalability,
growth, availability, and cost. Although fundamental, these
advantages are not easy to realize.
1.3 Challenges of Cluster Computing
There are a number of areas in which clusters are at a disadvantage
relative to SMPs. In this section we describe some of these
challenges and how they influenced the architecture we will propose
in Section 2.
Administration: Administration is a serious concern for systems
of many nodes. We leverage ideas in prior work [1], which
describes how a unified monitoring/reporting framework with data
visualization support was an effective tool for simplifying cluster
administration.
Component vs. system replication: Each commodity PC in a
cluster is not usually powerful enough to support an entire service,
but can probably support some components of the service. Component-
level rather than whole-system replication therefore allows
commodity PCs to serve as the unit of incremental scaling, provided
the software can be naturally decomposed into loosely coupled
modules. We address this challenge by proposing an
architecture in which each component has well-circumscribed
functional responsibilities and is largely interchangeable with
other components of the same type. For example, a cache node can
run anywhere that a disk is available, and a worker that performs a
specific kind of data compression can run anywhere that significant
CPU cycles are available.
Partial failures: Component-level replication leads directly to
the fundamental issue separating clusters from SMPs: the need to
handle partial failures (i.e., the ability to survive and adapt to failures
of subsets of the system). Traditional workstations and SMPs
never face this issue, since the machine is either up or down.
Shared state: Unlike SMPs, clusters have no shared state.
Although much work has been done to emulate global shared state
through software distributed shared memory [33,34,36], we can
improve performance and reduce complexity if we can avoid or
minimize the need for shared state across the cluster.
These last two concerns, partial failure and shared state, lead us
to focus on the sharing semantics actually required by network services.