Hey there,
I am having a bottlenecking issue with the way my network is set up. I am hoping for some advice and clarification on a couple things.
The network is setup like this.
D-Marc
Modem
Firewall
8 port Gigabit Switch.
Connected to this 8 port switch is:
our SBS server
1 16 port switch for a smaller network that does not access the sbs server
1 24 port switch that is located 250 ft away.
Connected to the one 24 port switch is another 24 port switch. So two 24 port switches connected in sequence.
Most of the bandwidth is being used by the two 24 port switches located at the front of the building and this is where I am having a bottleneck.
Mainly the only data that is being used is standard email, internet access, and file server access (all files are stored on the server). There is some larger files being transferred by graphics design and sometimes uploaded to FTP servers.
There is about 35 users with internet access. Is this bottleneck expected with this type of setup?
I am thinking my best option would to be running another cable from the server room 250ft up to the front wiring closet and connect both 24 port switches directly to the 8 port switch. Should this resolve the bottleneck?
I realize that the two 24 port switches should not be connected in sequence but it's the only option with the wiring available. Would i better off to put in another switch in the front closet and connected the two 24 port switches to a new gigabit switch?
Any other options to improve the bottleneck?
I have attached a photo.. Hi-Res to better show the setup.
Incase it helps Cisco PIX firewall,
Linksys 8 port managed switch
Dlink 16 and 24 unmanaged Gbps switches.
and Cat5e cable all around.
Your advice is appreciated!
Network Diagram
This Question has been solved and asker verified All Experts Exchange premium technology solutions are available to subscription members.
Experts Exchange has been collecting answers to technology questions since 1996…3 million and counting! If you have a question, chances are we already have your answer.
If you can't find the exact answer you're looking for, ask our exclusive community of 50,000 experts. You’ll get a personalized answer from a trusted professional.
Thousands of free tech tips, tricks, how-to’s and tutorials are available in our peer reviewed articles section. See for yourself how smart our experts are, no login required.
Save solutions to your questions, answers you’ve discovered through searching plus helpful articles in your personal knowledgebase for easy future access.
Access the answers to your technology questions today.
30-day free trial. Register in 60 seconds.
Members of the expert community talk about why the experience at Experts Exchange is different than what you will find anywhere else.
Try it out and discover for yourself.
30-day free trial. Register in 60 seconds.
Join the community of experts here and help other tech pros by answering question in your area of expertise. You can earn FREE access to all Experts Exchange's premium features and resources.
The bottleneck looks like its the single gig connection between the 2 24-port switches. I would try ethercahnneling a couple of connections between each switch to the 8-port switch if it's possible. But your best bet is bitting the bullet and fixing the cabling. You can also try QoS to give priority to traffic. This is of course if your switches can handle it.
Templar_m,
The bottleneck is experienced by the 35 users. This is connection to the internet as well as connection to the server. It is not all the time the bottleneck is noticed. It is most noticed at morning startup (8:00 am) and early in the afternoon, approximately 1-3.
The only managed switch is the 8 port gigabit switch. The rest are Dlink and are unmanaged.
Up to Date Norton Antivirus is on all of the PC's in the network.
clonga13,
I will be temporarily running a wire from one of the 24 port switches directly to the 8 port switch. This will mean that both 24 port switches will have a direct connection to the main switch. Only one of the switches used are managed I have changed the QoS priority level for the troubled interface.
hrodriguez761,
I have not tested the cable for attenuation. Is there a good way to test for this or do I need a specific tool?
Would I be better off connecting to one of the 100Mbps ports if attenuation was the case?
Hisham_Elkouha,
The lines that are from the switch to the round bubbles (indicating users) are to indicate a connection to the end user. There is only one line to the bubble but there is actually multiple lines to each network device. This being said, I dont think there is a loop in the layout. I did investigate STP though and will keep that in mind if I implement a second cable from the 8 port switch to one of the 24 port switches.
mikecr,
The switch is a decent switch Linksys SRW2008. I figured because there were segments that were attached to the switch that most of the work would be done by the other switches and the 8 port only handled internet/server traffic. If I only have 3/4 things connected to the switch should I still consider upgrading?
The Cisco PIX506e is the firewall that is inplace.
All,
One of the key issues with this is that outlook has difficulty connecting to the exchange server and starts slowing done or freezing usually resulting in a "Outlook is trying to connect to the exchange server" pop up baloon.
for testing the cable you can use ( Byte Brothers" RWC1000K ).
couple of points:
A.- I personally would not support 35 users out of a single 100mb port.
your best bet is an etherchannel if your equipment supports it.
B.- what is the model of your 24 port switches?
C.- If you connect each 24pt switch to the 8port switch directly you need to make sure STP is runing like Hisham_Elkouha said, otherwise you could end up having switching loops.
The Linksys with 8 ports isn't designed to handle that many users. I've deployed them in locations where I needed a couple of switch ports to add connectiivity but it is acting as a "core" switch on the network, seeing every packet going through your network and with that many users, like I said, it wasn't designed to handle that. One question I have would be are the users complaining about the speed of the internet or the speed of everything in general?
You have 35 users on one segment going through that 8 port switch to log into SBS, share files, get to the internet, and whatever else you have running on the SBS server.
hrodriguez761,
I will test for link attenuation soon.
A) 100Mb would only be used if attenuation is the issue. I will be continuing to use the 1Gb connections.
B) There is two different DLink Switches being used
C) I have ran a temporary cable from the other 24 port switch to the 8 port switch and removed the link between the two 24's. Both are now directly connected to the 8 port and no link between them. This should eliminate any looping. Performance has improved but there is still room for improvement.
mikecr,
It is the speed in general. My biggest issue is w/ our outlook connection to the exchange server. This is the most noticable as the slow down. I will be watching the task manager on the server when the connection error is occuring and there is not always high CPU usage. Internet is also slower than the 10 user segment.
There is two network cards on the server. Should I try connecting both NICS (Broadcom BCM5708C) to the 8 port switch and setting up Link aggregation? It looks like the NICs support TOE. Is this the same as NIC Teaming?
I am trying to avoid having to replace the 8 port switch but that may be a necessary option as we are still expanding our company. Is there any recommendations for an inexpensive backbone/core switch that can handle this type of load? Cisco/Dlink/Linksys suggestions please.
Another option I am trying to work out is connecting the SBS server directly to one (or both) of the 24 port switches. If I connect the server directly to one of the 24 port switches should I add the link between the two 24 port switches back?
One of the 24 port switches is an unmanaged switch. What can or needs to be done to implement STP
Thanks again for your help.
The optional layout involves directly connecting the server to one of the 24 port switches.
Current temporary setup is working well, it is only temporary because the cable run is not in a permanent location.
To test to see if the Linksys 8 port is the issue, take it out of the picture. Put the SBS and firewall onto the 16 port switch. Then make a run from from each of your 24 port switches back to the 16 port switch making sure that both 24's are not hooked to each other so you don't get a spanning tree loop. This will utilize the 16 port and 24 port switches only. Test your connectivity then and see if it is improved. If the 16 port is a manageable switch, make sure the speed/duplex settings match on the server and the switch.
This test will see if you really need to put in any type of core switch. If it works good than you can just leave it alone. Security is the same throughout the network since the Pix doesn't change.
I have tested with the 8 channel in and out of the picture and I find I have a bit better success with it in. I think this may because of the volume of traffic that is being used on the 16 port currently.
The 8 port is a manageable switch and I am testing out Link aggregation with 2 connections to the SBS server. Things are going pretty well so far.
I am going to be getting some more cable runs installed from the Server Room to the wiring closet with the two 24's.
I am also considering upgrading to manageable switches instead of the unmanaged 24 port switches.
Does anyone have any comments or concerns with the link aggregation?
On the SBS server we have two identical broadcom NICs that I have set up the teaming on.
There are two types of configurations, load balancing and redundancy only. If you choose load balancing, you will effectively increase the load from the switch to the server. If this is a bottleneck point, you should notice some increase in speed. If you don't, then the issue may be a workstation or server configuration issue.
If you wish to move to a different core switch, I would recommend Either a Cisco 3560 or 3750, or, a DELL 3524, with the DELL being the cheapest.
I have made a few changes to the network but have not Upgraded all the switches at this point.
I have installed more wiring between the core switch to the front 24 port switches. I've also upgraded to Cat6 Cable instead of the Cat5. We are getting 1000Mb connection on the Cat6 cables.
There has been many alterations on the server also, such as adding an additional hard drive for Page Files and Log files (This is to reduce the IO on the main hard drives). Removing stale RUS data, updating storage drivers, etc.
I have turned cached mode on for the heavy email users and this has made some improvements if the connectivity is lost, they still have access to email and are only displayed the 'trying to connect' pop up message. Good improvement.
I have monitored higher activity ports and addressed the computers they were connected to. Installing anti-spyware and anti-malware. I've considered adding them to a VLAN to see if lowered the impact of the chatter. Very little or no spyware/malware has been detected.
Next step is going to be switching antivirus to a server version instead of having local installs on all the machines.
Business Accounts
Answer for Membership
by: Templar_mPosted on 2009-09-04 at 12:59:42ID: 25262860
Where do you experience the bottleneck? Is it traffic to/from the internet or traffic to/from the server?
Do you experience the slowness all the time or certain periods of the day?
What kind of switches do you have? Managed or unmanaged? Do you have an antivirus on your network?