My scenario:
CentOS 6.5 on VirtualBox, Bridged Interface with WLAN card(Cable connected)

My Configuration:
ifcfg-eth0 configured to use static IP, nameservers are that of google's.

Problem:
While I was updating, I noticed that from my full download speed, the speed slowed down to a few bytes and eventually leading to loading of new mirror. This worked fine for 3 to 4 seconds and again slowed down. It took me a lot of time to download a file which otherwise would have taken just 2 minutes.

I am pinging google.com and it seems that ping is fine for 3 seconds, then stops, then again starts after 5 to 6 seconds, then stops and this continues.

When I ping from my host, it is normal and connectivity is like it should be. From VM, its intermittent. Simultaneous pings and the same result.

Need your help!

I am still facing this issue. Can someone suggest what I need to look at? I have done a fresh install, reconfigured static IP and related information. The problem persists.

##Downgraded to CentOS 6.4

Are you running this in NAT or bridge mode (in VirtualBox)? If bridging, this sounds like an ARP timeout issue. If NAT'ing, it may be a bit more complex.

Its in bridge mode(Cable connected). It pings for 3-4 seconds and then stops for 3-4 seconds and then starts and stops. Due to this, when I am upgrading my repositories, it exhausts all the mirrors because of this and in the end it dies in between.

Try switching to NAT mode and report.

If NAT mode works consistently, it is likely an ARP issue when trying to juggle two MAC addresses on the same WiFi client attachment, or the WiFi equivalent of a "port". Non-WDS access points struggle with this and it is not supported by the WiFi spec.

It sounds like Layer 2 is where your problem is happening. Let us know if NAT makes a difference; if so, I have a couple more things to try to verify and possibly mitigate that issue.

NAT indeed works perfectly and I have tested it before. Actually, I used to use NAT before, but switched to bridge connection to run some scans and tests.
Let me know the other tests you are referring to. I will give those a try.

Open a second console window to broadcast gratuitous ARP and execute the following (omit 'sudo' if logged in as root; you'll need root permissions to open the raw socket) and replace "<IP>" with the IP address on the VM (in bridged mode):

While this is running, test connectivity for a bit using another PTY. Let me know if you see the same intermittent connectivity, or if it improves. You might also want to keep an eye on the connectivity of the host system.

This morning I booted up virtual CentOS and did a "yum update" and to my surprise, it updated without any problems. I didn't even tried "arping" as you suggested.
After upgrade, I tried "arping" and I couldn't pin point anything. Connectivity seems flawless(since I was able to update without much effort).
I am not sure how to reproduce it now. And I can't say for sure that this will not happen again.
What do you think Solarisguy?

Ok, so the problem started happening again. This time I took a snap of it. In the capture file that I have attached, we can see that there is a loss stated below and this can be verified by the numbers of ICMP packets.
At the same moment, I was pinging my gateway from my physical host and there was no loss. This explains why internet sites are giving problems. My gateway itself is dropping packets and the way I see it, it's for no reason.
Any help will be appreciated.

Attached Thumbnails

 

What are the details on the gateway? I assume it's a SOHO consumer WiFi router?

Its WAG54G2 Wireless ADSL2+ Gateway.

This is more than likely a problem with the wireless driver on the PC, or in the wireless stack on the router.