Research and evaluate the range of issues that are likely to occur within a converged network environment.

A converged network is defined as : A single network with the capacity to carry a combination of data, voice and video traffic.(Global Knowledge , 2010). Data and Voice/Video have very different needs from the network, data needs acknowledgement that data has arrived i.e. (TCP) , Voice/Video don't need all the data but need data at the right time to output sound/images in real time i.e. (UDP).

Why do businesses want a single network?

According to “How does network convergence support a business strategy?”, Cisco, (2003) pg9) the main benefits of a converged network with examples in Santander are:

1. Cost reduction and management - WAN network costs are decreased because two i.e. (one main and one backup connection) lines, instead 100's of phone lines via PABX , can be used between office locations, making it easy to switch between ISP's (Internet Service Provider) forcing them to compete with each other.

2. Property and facilities utilisation and management - Network management responsibility is passed onto Santander. Data, Voice and Video network traffic now have to be routed, sorted and managed internally.

3. Organisational agility - IT staff can now service changes and problems on site were before one would have to wait for a ISP technicians.

4. Employee productivity - office works have an array of tools that are now integrated allowing workers to share screen informations while talking over video conference improving employee productivity. However According to ZDNET(2010) 90% of security threats come from employees.

So there are two main issues from these benefits which are, X now needs to manage network traffic and security.

How do network administrators manage traffic?

The tool used to manage these different network needs is known as QOS (Quality of Service) which provides the ability to provide different data flows. There are three main factors in QOS → Delay i.e. (time taken by packet to destination from A to B) , Jitter I.e (time to destination between packet A and B) and bit rate.

The bellow table summarizes the differences between Video, Voice and Data:

	Video	Voice	Data
Example	Video Conference	VOIP phone	FTP(File transfer protocol)
Packet size	+-256 – 1518 bytes	+-180 bytes	+-512 I.e.(greater if reset MTU (Minimum transmission unit))
Maximum Delay	If video stream up to 3 minutes if conference +-150 ms	+-150 ms	Not greater than packet TTL (Time to live
Loss Sensitive	No,but effect quality of video.	No, but effect quality of voice.	Yes, Corrupting files
Jitter Sensitive	Yes,effect quality of video.	Yes,effect quality of voice.	No
Bit rate	64kbps	16 kbps	Depend on file size
Transport	UDP	UDP	TCP
COS and/or TOS (Precedence bits)	Example: Bits 100 → 4 (High priority)	Example: Bits 010 → 2 (Medium priority)	Example: Bits 001 → 1(Low priority)
Table: QOS - Video, Voice and Data (Cisco, 2010), (Wikipedia, 2010)

Queues are used to prioritised traffic with Qos markings. In the DSCP (differentiated services code point) standard within Ipv4 there are two fields in the IP header, COS(Class of Service) → ISO layer 2 marker, and TOS (Type of Service) → ISO layer 3 marker.

In the LAN environment COS is used to prioritise traffic that is going to a workstation with PC and VOIP. These two types of traffic may be carried over two separate VLAN´s, with many small UDP VOIP and some large TCP application packets. A PC application might mark its own packets with COS values e.g. (bits 101 → 5) that are higher priority than the VOIP packets thereby effecting VOIP packets and reducing Voice quality.

In a WAN environment where COS values are replicated in TOS, WAN connectivity and data rates have to be guaranteed through SLA (Service Level agreements) as QOS can only guarantee QOS up until the network edge provided there is no LAN congestion.

The highest priority (bits 111 → 7) are reserved for routing packets. Why?. Without routing packets routing table would never be able to be built i.e. (Unless you using static routes) making routing impossible!!

EXAMPLE PROBLEM: The above table classification classifies data as the lowest priority but this is not always what is needed. In X daily batch files are sent with financial information that are of higher priority than Voice packets. The application that sends these batch files are allowed to apply there own COS priorities.

In a big bank like X there are hundreds of different types of applications which can't all be classified into only 6 categories. A alternative classification system of DSCP extends the TOS bits by three adding more room for classification.

How do network administrators manage security?

Moving to a single network gives more network management control to X enabling them to adapter quicker to changes in the business environment. Security has become an important topic, especially for banks.

LAN Switch-ports need to be configured as access, voice or trunk ports but this still doesn't stop VLAN hopping i.e. (the port might be configured as dynamic) . Switch-port violation configurations need to be set but can result in more calls to the IT department because interfaces have been disabled by security violations reducing employee productivity.

Other security mechanisms like dot1x authentication might not be possible because the application batch server might not support dot1x requiring forced-authentication. In the WAN environment VLAN trunks are not encrypted. This is seen as the job of other encapsulations such as IPSEC which provides a complete security tool set.

Conclusion

As more management and control of the networks moves over to X the list of services that can be supplied to employee workstation grows. The network becomes more adaptable to changes in the business network environment instead of the ISP network environment. Resulting in a network that that meets user expectation more closely.
Security and QOS have become key issues but have also created new ways of improving worker productivity through VOIP and Video Conferencing. Security was a huge unmanaged risk in the old network model with service providers out of touch with end users such as X.
X now has a network that is more important and services there needs more than ever. The single network helps X be a better bank.

Evaluate the range of issues that may occur in the routed environment, including a detailed exploration of the issues affecting router performance when EIGRP, OSPF and BGP are used.

At X EIGRP, OSPF are used for countries or regional offices. BGP is used between Countries and high bandwidth WAN links.

Bellow is a Summery of the differences between the routing protocols (Cisco,2010):

	OSPF	EIGRP	BGP
Protocol type	Link State	Advanced Distance Vector	Path Distance Vector
Classless	Yes	Yes	Yes
VLSM	Yes	Yes	Yes
Automatic Summarization	No	Yes (Possible to disable)	Yes
Size	Large	Large	Very Large
Convergent time	Fast	Very Fast	Slow
Metric	Cost	Composite Metric	Path Attribute
Proprietary	No	Yes	No
Complexity to configure	Medium	Low	High
Boundary separation	Areas	Summarization	Autonomous Systems
Administrative distance Internal	110	90	200
Administrative distance External	Na	170	20
Table - Summary EIGRP, OSPF, BGP

Why so many routing protocols in one network?

EIGRP and OSPF are primary used as internal gateway protocols within and an autonomous system i.e. (Where one group of administrators have control). X primarily uses OSPF but there are sum sections of the network that use EIGRP because of old infrastructure from mergers and acquisitions.
Santander prefers OSPF because it is non-proprietary i.e. (cheaper) and can be used in large networks serving the same purpose i.e. (To provide routing in large network) as EIGRP. Also not all network related equipment within X is CISCO. Santander could be loosing in convergence speed, but OSPF can be custom configured to converge as fast, or even faster, than the default EIGRP configurations.
BGP is slow in convergence and has large routing tables i.e. (which are made smaller through summarization) but is a robust protocol that is used by the internet. BGP does not automatically discover neighbours through advertisements such as OSPF or EIGRP. Neighbours need to be manually configures for BGP, then information is automatically exchanged between these neighbours.
A major danger with BGP is the “clear ip bgp *” which can cause routing tables that are 100MB's big to be downloaded not to mention that BGP is used to connect high bandwidth WAN links resulting in loss of connectivity to whole countries.
Main and backup ISP's are used to connect regions via WAN connection types such as Frame relay or P2P. BGP is well supported by ISP's where EIGRP or OSPF are not. ISP's use BGP because of the size of there networks and large routing tables that would demand more processing power and bandwidth if they had to converge or send as many neighbour advertisements as EIGRP and OSPF. Also ISP's are often measured by there SLA agreements and can't afford to lower connectivity/service levels for constant topology changes.
To summarize the reason we have different routing protocols like OSPF, EIGRP and BGP is to serve very different business needs. OSPF and EIGRP are used for small to large sized networks which require fast and frequent convergence and topology changes. BGP is used for huge networks which require robustness and have few topology changes.

How the performance of a router is effected by its protocol?

There are 3 factors that effect the router performance of a protocol. (1)The size of routing table and topology tables i.e. (effecting memory) (2) Amount of communication between nodes i.e. (effecting bandwidth) (3) Algorithm handling of Topology changes i.e. (effecting CPU utilization).(Cisco,2010)
To summarize the Memory, Bandwidth and CPU utilization of OSPF, EIGRP and BGP:

	OSPF	EIGRP	BGP
Memory	Medium – Store all link states	Low – nodes learned by neighbours.	High – Big Routing tables
CPU	Medium– Dead Timer 40 sec but more CPU cycles to maintain link states, SPF (Shortest path first)	Medium – Hold Time 15 sec, DUAL( Diffusing Update Algorithm)	Low – Hold timer 180 sec.
Bandwidth	Medium - Hello (10 sec), DBD, LSR, LSU, LSAck.	High - Hello (5 sec), Update, Query, Reply, , ACK	Low - Keep-alive (60 sec)
Comment	Optimize summarization at border routers and stubby areas.	Optimize router with summarization.	Optimize router with summarization.
Table – Router Performance EIGRP, OSPF, BGP

Looking at the above table one can see OSPF and EIGRP come out the same with an average of medium while BGP coming first with the lowest utilization. So why doesn't everybody just use BGP?

The main reason been the slow convergence and those huge routing tables. With convergence of services over one link this would mean a large business would have to go a full 3 min or more with no VOIP(Voice over Internet Protocol) telephones or applications and internet. In Santander, BGP routers are the gateways into networks that are connected to large bandwidth connections on heavy duty routers such as the CISCO Catalyst 6500,which are not cheap, including a cost factor.

OSPF seems like an obvious selection over EIGRP because it has a medium performance like EIGRP but is not proprietary making it cheaper. However EIGRP is easier to configure if you in a CISCO environment and therefore requires less labour input bringing down costs. However Santander is not in a CISCO only network environment making OSPF the preferred protocol.

Conclusion

OSPF, EIGRP are overall performance wise very close but a business needs to ask its self “Am I working in a CISCO environment?” and “Do I have the manpower and skills to use this protocol?”. Making the deciding factor between EIGRP and OSPF the business environment.
BGP which is really the only established path distance vector protocol out there is a specialized protocol that specializes in long distance travel across Autonomous Systems and big networks. Its Robustness is its biggest weakness when it comes to small to medium sized networks making it unsuitable for a internal gateway protocol, however its biggest strength is also its robustness when it comes to large networks like the internet.
So one should choose a routing protocol based on the situation.”Proprietary or non proprietary”, “Manpower or no manpower” and “Robustness vs speed and convergence” to get the best value out of routers.

Evaluate the impact of ‘change’ on the management of a corporate network infrastructure.

At X the words “Change management” are read and heard at various times during the day in the from emails of scheduled changes, change management procedures, to the weekly change management committees. Why? What are they trying to achieve?
The objective of Change Management in this context is to ensure that standardized methods and procedures are used for efficient and prompt handling of all changes to controlled IT infrastructure, in order to minimize the number and impact of any related incidents upon service. (Wikipedia, 2010) – So these procedures make changes in a more efficient way while minimizing the impact to Santanders services which is what they provide to there customers.
A summery Santander Change Management and Procedures is shown bellow:

ITIL Model	Procedures for Change Management	Tools
Change Management Incident Management Problem Management Release management Configuration Management	Filtering changes Managing changes Change Advisory Board (CAB) and Emergency committee Reviewing and closing of Requests for Change (RFCs) Management reporting	Remedy, Tacacs TFT server, Wiki website Spectrum, Microsoft Visio, Ping, Traceroute, Cisco show commands, CACTI
X - Models, Procedures and tools - (Troubleshooting and maintaining Cisco IP networks (TSHOOT) foundation learning guide, Cisco Press, pg 5)

X uses the ITIL model – which is a IT service management framework describing best practices that help provide high-quality IT services that are aligned with business needs and processes. (Ranjbar A., 2010)
The above ITIL model implemented at X is a Structured driven approach with the impact been high availability and reliability. Changes are carefully planned and design and controlled by the CAB and later reviewed to see if the changes have completed there objectives i.e. (To provide better Service). Tools such as Remedy change management are linked to ether new implementations or incidents that have become problems and scheduled for a change i.e. (A reactive approach, Filtering and Managing changes).
The “Network Monitoring Department”, uses a preventative approach, with tools Spectrum, TACACS, CACTI. Investigation of alerts from these tools using command such as “show logs” i.e. (Requesting changes if needed). Commands such as “show controller”, “show interface X” are used to confirm that changes have been implemented. i.e.(Review of changes)



The below table shows how the ITIL business needs are linked to processes. Reducing the impact of changes and increasing service levels:

Needs	Processes
Resilience	Redundancy – Duplication of lines, servers, etc.
Reliability	SLA (Service level agreements), MTBF (Mean time between failures), MTTR (Mean Time To Repair) reports by management.
Maintainability	Monitoring, Auditing, Backup and Roll-backs by staff and departments.
Serviceability	Telefonica, Indra and Orange and reports on compliance with SLA`s. i.e. (management reports from “Table - Models, Procedures and tools”)
Security	Authorisation, Authentication and Accounting of Routers, Servers and Systems of users.
Santander – Business Needs and processes – (Wikipedia – Change Management, 2010)

Natural Disasters

However the use of ITIL and tools along with change management doesn't mean that the network never has interruptions. Service providers have unplanned maintenance, routers simply crash after running perfectly for 5 years or natural disasters cut fibre optic lines cutting off whole countries. i.e. (interrupt or incident driven approach)
The result been that change management ,can't guarantee a 100% up time.
e.g. (An earthquake in Chile disconnected the network )
NOTE: Regular incidents which become problems become part of the change management process. Incident management is a solution to interrupts and real time problems however still does not guarantee 100% up time.

Cost Factor

In general X aims to reduce the MTTR to less than 48 hours and this is reflected in the SLA agreements with service providers such as Y and Z.
However this dose not always make economic sense. The marketing department might calculate that country x has more potential for profit than country y reallocating resources from country y to country xe.g. (Swapping a higher capacity router in country y for a lower capacity router so that country x can have more higher capacity routers) . Resulting in changes in y been slower because of lack of funding.
The result been better Service is not always the only objective.

Political factor

Department x is going to be phased out because they are not part of X new strategy so MTTR (Mean Time To Repair) is much longer for department x. i.e. (Why install a new router when you going to remove it in 6 weeks!)
The result been MTTR (Mean Time To Repair) is not always minimised because the objective is not to improve service in department x.

Conclusion

Change management and ITIL is not a flawless framework that guarantees perfect services and 100% uptime but it does manage changes and has both a structured i.e. (change management) and interrupt driven i.e. (Incidents and problems) approach. Which does minimise impact therefore increasing service levels.
Natural, Political and Cost factors might cause services to be effected. X might simply not want to improve service because it is not part of there objectives. X can't control the environment, resulting in real world problem solving and problems that ITIL can't solve so a common sense approach is still very relevant.
Business objectives and common sense still matter over ITIL.