Wednesday, December 30, 2009

Modes Of Ethernet Over MPLS (EoMPLS)

An Ethernet Pseudowire allows Ethernet packets to transport over MPLS cloud. By using this service customers simply extend their local area networks without loosing the information. The spanning tree will work and end to end connectivity will be on same subnet. Otherwords, we can say it a virtual leased circuit. Customer end to end connectivity is aka as Emulated Service which operates over Pseudowire which further operates over Packet Switched Network (MPLS Network). Entire end to end communication reference model is depicted below.

Figure 1

Modes Of Ethernet Pseudowire
An Ethernet Pseudowire operates in two modes: raw mode and tagged mode. In tagged mode, each frame must have 802.1q tag and that tag is meaningful to the local and end point router. It should be noted that if the VLAN identifier is modified by the egress PE, the Ethernet spanning tree protocol might fail to work properly. If this issue is of significance, the VLAN identifier MUST be selected in such a way that it matches on the attachment circuits at both ends of the PW. It means the identifier or vlan tag should be used same not different. This mode use Pseudowire type 0x0004. Every frame sent on the PW must have a service-delimiting VLAN tag (Different vlans for different customers). If the frame as received by the PE from the attachment circuit does not have a service-delimiting VLAN tag, the PE must prepend the frame with a dummy VLAN tag before sending the frame on the PW.

But in case of raw mode, tag may or may not be added in the frame and is not meaning to the end points. Though the frame is forwarded transparently. This service corresponds to PW type 0x0005. If an Ethernet PW is operating in raw mode, service-delimiting tags are NEVER sent over the PW. If a service-delimiting tag is present when the frame is received from the attachment circuit by the PE, it mUST be stripped from the frame before the frame is sent to the PW.

Why Customer Needs Extended Lan Services
1. In case of Data Center, customers having geographically seperated DC and want to replicate the storage over Fiber Channel IP aka FCIP. FCIP works only on same broadcast domain.
2. Need to run dynamic routing protocol between the sites.

Click Here To Read Rest Of The Post...

Monday, December 28, 2009

PE-CE BGP Link Protection - Testing

In my previous post I described the loss of communication in PE-CE BGP routing. The same could be overcome by using link protection. The testing of PE-CE link protection document is available. Click here to read the full story.

Click Here To Read Rest Of The Post...

Wednesday, December 23, 2009

PE-CE BGP Link Protection In Case Of Dual POP

Customers always complaint about the connectivity loss in case of dual connectivity from service provider. I have seen this issue in many cases when BGP is implemented as PE-CE routing protocol. BGP default behaviour, after a PE-CE link failure nad its detection in routing, all the BGP learned routes removed via the failing link. BGP runs the best path algorithm for the failure routes and then selects the alternate path. The local labes are removed and the same message of withdrawal is sent to all the BGP neighbors. After the withdrawal message the best path algorithm runs and new prefixes are switched to the alternate paths. This is time consuming process, normally take less than 180 ms. But sometimes the applications are so critical and the customer could not bear even the loss of seconds. For more click on Read More!

For this existing problem of long convergence Cisco has introduced new solution of link protection. The MPLS VPN PE-CE Link Protection feature reduces time by sending the broken link’s traffic over a backup path instead of waiting for total network convergence. The feature accomplishes this by maintaining the local label (for 5 minutes) for the prefixes which are switching from the failing local path to the backup path. Because the label is not freed as had been the usual practice, forwarding can continue to take place, but now to the alternate (backup) path chosen by the bestpath algorithm. Thus, the local label has been applied in place of the failed BGP bestpath label; (this is sometimes called “label swapping”). Traffic thereby is locally restored, while the network propagation of the BGP withdrawal messages takes place. Eventually, the egress PE router converges and bypasses
the local repair
The given feature is enabled under vrf and address-family of particular vpn on primary PE
protection local-prefixes

Click Here To Read Rest Of The Post...

Monday, December 21, 2009 started

I have booked the new domain My existing blog traffic is already redirected to new domain.

Click Here To Read Rest Of The Post...

Saturday, December 19, 2009

Tool to design and configure networks

Businesses rely on their network infrastructure to connect their employees together, connect them to other businesses, and more generally, connect them to the Internet. Networks have grown more intelligent and more complex and users have become more technology savvy. Network devices have become a commodity, supporting increasing functionality. Users expect more.

Lucipa believes that with the right tools, the task of configuring a network can be largely automated and streamlined. Lucipa's software, SpiderWeb, will generate the required configuration of devices based on industry best practices.

SpiderWeb solves the problems of configuring increasingly complex features in networks; it can create compliant configurations by providing a user intuitive software. SpiderWeb allows engineers to create a diagram of a network, assign common parameters to all devices, input device specific parameters, and get the configurations of all devices in the network.

Spend less time creating configurations and spend more time with your customers

You can design the network, create reports, and create all the configurations before starting implementation. SpiderWeb helps during the planning phase, design, phase, and implementation phase. Actually, companies can perform these steps 5 to 10 times faster than when using manual methods.

The system produces the web reports, policies and having a very good expert system.

Sources:- Lucipa Website

Click Here To Read Rest Of The Post...

Friday, December 18, 2009

Fragmentation choke bandwidth

Link was working fine with ip traffic and getting the proper load. But as soon as the l2tpv3 traffic was added customer won't able to the add the more traffic on configured link. After deep analysis, we had seen the problem occured because of fragmentation issue which was being caused by l2tpv3 consequences router ip input was increasing drastically. The solution of the problem is either to set the path mtu or increase the mtu in the whole path.

Click Here To Read Rest Of The Post...

Thursday, December 17, 2009

How to calculate end to end transmit power in fiber

Calculation of power budget in fiber deployent is one of the biggest challenge. If the power is more or less than the power budget, it will bring lot of problems like throughput, drops and many more. One must ensure that the end to end link is working up to the level and able to delivery according to the commitment and design. To ensure these parameters, one must know about how to calculate the end to end power budget. Power budget is nothing but it provides the margin that compensates for variations of transmitter power and receiver sensitivity levels.

Power Budget = Minimum Transmit Power - Minimum Receiver Sensitivity Level

Minimum and maximum receiver and transmit levels could be obtained from the product specifications.
With this we need to calculte the total loss caused by the equipments in end to end deployment. Click here to see the values of losses w.r.t to equipment.

End to End Loss Calculation = (Fiber attentuation*km) + (Connector attenuation *Number of connectors) + (Splices * Number of splices) + (Passive devices*Number)

The above mentioned calculations are basics but the question comes in mind that how much power should one point transmit and another point will receive. Whenever a starting point will transmit some power the same power doesn't receive by the another point; This is because of losses in the path. The another point is that the power received by the end point should not exceed its maximum receiver sensitivity. So we can use the given formula to calculate the input power to link.

Input Power = Maximum Tranmit Power - End to End Loss Calucation < Maximum Power Receiver Sensitivity
example will help to understand the whole calculations w.r.t to the live deployment:-

The minimum optical transmitter launch power is –12.5 dBm, and the maximum optical transmitter launch power is –2 dBm at 1310 nm. The minimum receiver sensitivity is –30 dBm, and the maximum receiver sensitivity is –3 dBm at 1310 nm. Total fiber stretch is of 3 Km with 3*fusion splices and 2*patch panels. We can ad the connectors also but in this example I am using that.

First calculate the total power budget which means how much maximum power we can allow for the link. As mentioned above in the formula we need to add the values.

Power Budget = Minimum Transmit Power(-12.5) - Minimum Receiver Sensitivity Level(-30)
= -12.5 + 30
= 17.5 dB

Now calculate the losses

This is the MMF, assume graded index 50/125 = 3 Km * 0.7 dB/KM => 2.1dB

2* Fusion Splices = 2* 0.02dB => 0.04dB
2* Patch Panels = 2* 2dB => 4.0 dB

Total Fiber Loss in 3 Km = 2.1 + 0.04 + 4 => 6.14dB

Now we need to calculate the input power:-

Maximum Transmit Power(-2) - Total Fiber Loss(6.14) < Maximum Receiver Sensitivity -2-6.14 => -8.14dB < -3dB

Click Here To Read Rest Of The Post...

Tuesday, December 15, 2009

Values which can affect power budget of fiber

See the previous posts of Fiber Splicing and Difference between Single and Multimode Fiber.
To calculate power budget we should know the fiber attentuation characterstics, Component loss values and reference margin values. Once all the values are known one can easily calculate te power budget.

Fiber Attenuation

Insertion Loss With Respective Components

Loss Margin

Click Here To Read Rest Of The Post...

Tuesday, December 8, 2009

Types Of Fiber Splicing

What does happens when fiber cut occur? Someone need to fine the location of cut and restore. The restroration technique of fiber or merging the two fiber ends in one to increase its length is known as splicing. It is done in two ways

1. Mechanical Splicing
2. Fusion Splicing

Every splicing add some looses in the end to end calculation. For fusion splicing a loss of 0.02 dB is measured and for mechanical loss of 0.75 dB is measured. If it is not done in good manner it could increase the losses consequence drop in the data rate. The best way to check the total number of splices in end to end circuit is to see the OTDR (Optical Time Domain Reflector) report which helps to find the problems in the fiber.

Click Here To Read Rest Of The Post...

Monday, December 7, 2009

Single Mode and Multimode Fiber

I am always bit confused about single mode and multimode fiber cables. The differentiation between them is not so tough to understand. The only difference is the core part.
The core is the highly refractive region of the optical fiber from which the light is trasmitted. According to the ITU the diameter of the single mode fiber should fall between 8 to 10 micron meter and 50 to 62.5 micro meter of multi mode fiber cable.
Now how to remember which cable is used small or large distance. It's also very easy. Think single mode fiber cable is having a very small core where in only a light can transmit, it means a monopoloy of light is there and it could travel upto to a large distance. The multimode fiber cable is having a big core which means more lights could travel in the big core, a quarrel always happen where more than one comes, so no one can survive for a long period of time. Hence MMF cables are used for short distance.

Click Here To Read Rest Of The Post...