Exam ANS-C00 topic 1 question 138 discussion

If packets are over 1500 bytes, they are fragmented, or they are dropped if the Don't Fragment flag is set in the IP header. We want the packets to be fragmented and not dropped. So Asnwer : A

for me its 'A'

I initially went for D, but there is nothing in the question that states that jumbo frames must be used end to end. The simplest way to resolve the issue is therefore A

The answer is certainly A for a few reasons: 1. Not all DX connections support jumbo frames (To check whether a connection or virtual interface supports jumbo frames, select it in the AWS Direct Connect console and find Jumbo Frame Capable on the Summary tab. - https://docs.aws.amazon.com/directconnect/latest/UserGuide/set-jumbo-frames-vif.html) 2. Not all AWS resources support jumbo frames (same link as above - All EC2 instance types support jumbo frames except for C1, CC1, T1, and M1) 3. People who went straight for D are trying to fix a simple problem with a complicated, expensive and time consuming solution.

To support the jumbo frame..DX is required from on-premise to aws..If you use vpn to connect a vpn, it only supports 1500MTU. ABD are fine. the problem is which is the optimal course of action... A mean the network does not support jumbo frame and the packets will be dropped if the router see the header has the "do not fragment" flag. So choosing A, the packet will be fragmented to 1500 and slow down the performance. The outcome of B is similar as A. Both answers are fallback actions to 1500MTU... If you do not want to take the advantage of the jumbo frame...then choose A and B. Otherwise, you should take D. It is multiple choice question so you cannot choose A&B. So it should be D.

Ans = D; remember that the question says "What is the optimal course of action?" the most optimal is to have the connection through a Dx so that you can take advantage of the Jumbo Frames Feature. why would you limit you capability if all your nodes already support 9001 MTUs?

Both A and B would work. But I would suggest B since fragmenting packet affect latency. Better fix the problem than have a workaround ..

A In the question, it didn't say direct connect is required from bandwidth perspective. The issue is you can't access AWS from on-premise network. It's unreasonable to upgrade to direct connect from this.

I think "A" is the correct answer https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/network_mtu.html

Jumbo Frames are only supported in Direct Connect (not VPN). So for me D is the only choice

Regular packets do not have the DF bit set Also, fragmentation is aweful Lowering the MTU for your network is the way to go, despite such task being not so easy But hey .. when you do crap, you have to pay the price

And this is why jumbo frames will be fragmented, unless "Do not Fragment" flag is attached. Then they will be dropped. Am i right ?

The servers should use PMTUD to discover MTU for different destination so it can continue using jumbo frame inside the company and smaller packet size for reaching the AWS. I guess the question assumes that PMTUD is not enabled or available. In this case, both A and D will work. D say you must upgrade to DX that might be too extreme. A?

B you cannot traverse using MTU to AWS, only MTS is supported within your VPC (peering also is supported between VPC's)