06-20-2013
Parkinson's law: need expands to consume excess resources.
Companies like Akamai make a good living ensuring your static web hits are filled from relatively local cache servers.
Good architectural design has to deal with:
- having a soft saturation, so throughput goes up to saturation and then excess load is shedded in a least-lost-value basis, like newest clients lower in priority than older clients (deeper into transaction process).
- avoiding negative saturation behaviors like overloaded Ethernet, which actually slows down due to collisions creating lost time on wire. Positive saturation behavior means the higher the overload, the more efficient the process. Requests can be sorted to they have higher locality of reference. Sometimes, requests for the same file can be mbone multicast as one. Sorting by disk position means shorter seeks.
- flow control mechanisms allow services beyond capacity to be queued for eventual fulfillment, but service cancellation is quickly forearded to the server. The bad behaviors are thing like sending service requests every n seconds until a reply is received, consuming precious bandwidth and cluttering the server with cancelled, redundant, prior requests. Some routers can stifle keep-alive traffic, say from tcp connections of queued services, so they do not drag down net speed.
9 More Discussions You Might Find Interesting
1. UNIX for Dummies Questions & Answers
I'm on T1, is there any way i can check my acctual bandwidth?
Thank you all (3 Replies)
Discussion started by: solvman
3 Replies
2. IP Networking
i have two questions actually...
i need to block certain ports with openbsd and PF in a large lan, the firewall is supposed to be a router between the internet and the first lan switch.
first of all, would this work at all in theory?
second, i tried doing this a few days ago at a huge lan but i... (2 Replies)
Discussion started by: nocturnal
2 Replies
3. IP Networking
Hello,
I was wondering how one would go about regulating bandwidth of a server running RH Linux 8.0 and Apache. I am running a webserver, and want only one person to be able to download one file at a time. So, they can't click about 10 files and soak up lots of bandwidth. Plus, I would like to cap... (2 Replies)
Discussion started by: Phobos
2 Replies
4. UNIX for Dummies Questions & Answers
Hi,
I'm looking for some way to bring up in a shell window a realtime (or something that updates at 10 second intervals or whatever) bandwidth monitor.
I just want something that shows me how much kbps is going in and how much is going out of each interface. Is there something that might... (2 Replies)
Discussion started by: d11wtq
2 Replies
5. UNIX for Advanced & Expert Users
Hi Gurus,
Need to have a way to monitor Bandwidth utilization on Linux servers, running squid .
Have worked on 3rd party monitoring tools like Bandwidth d, Nagios etc.
But we are working to find out a way to monitor this through Sitescope, for which we need to find a file where the... (4 Replies)
Discussion started by: Crazy_murli
4 Replies
6. IP Networking
Hi,
1-What is bandewidth ?
2-How to calculate ?
3-How to measure ? Any free software to measure it ?
Many thanks. (4 Replies)
Discussion started by: big123456
4 Replies
7. IP Networking
Hi all
I have been running iptraf on a linux box for a few months monitoring bandwidth utilization realtime from mac address with port mirroring.
Now I want to graph these bandwidth utilization. Is MRTG the right software?
Any ideas? (4 Replies)
Discussion started by: palm101
4 Replies
8. Linux
Dear Engineer,
Can anyone please inform me some application name like (CBQ) to control network bandwidth and please provide me the url to configure it.
With Best Regards,
Md. Abdullah-Al Kauser (0 Replies)
Discussion started by: makauser
0 Replies
9. Red Hat
Dear Engineer,
Can anyone please inform me some application name like (CBQ) to control network bandwidth and please provide me the url to configure it.
With Best Regards,
Md. Abdullah-Al Kauser (1 Reply)
Discussion started by: makauser
1 Replies
lockd(1M) lockd(1M)
NAME
lockd: rpc.lockd - network lock daemon
SYNOPSIS
graceperiod] listen_min_backlog] timeout] [nthreads]
DESCRIPTION
The lockd utility is part of the NFS lock manager, which supports record lock and share lock operations on NFS files. See fcntl(2) and
lockf(2). provides two functions:
o It forwards locking requests for NFS mounted file systems to the lockd on the NFS server.
o It generates local file locking operations in response to requests forwarded from running on NFS client machines.
State information kept by the lockd about these locking requests can be lost if is killed or the operating system is rebooted. Some
of this information can be recovered as follows. When a server recovers, it waits for a grace period for all NFS client-site to
submit reclaim requests. Client-site is notified by the of the server recovery, and promptly resubmits previously granted lock
requests. If a fails to secure a previously granted lock at the server site, then the sends a to the process holding that lock.
Administrators can make changes to the startup parameters for by logging in as root and editing the file (see nfs(4)). Administra-
tors can also make changes by using the command.
Options
recognizes the following options and command-line arguments:
Deprecated in favor of the
parameter in the NFS default file, which is equivalent. Specify the number of seconds that all clients (both NLM
and NFSv4) have to reclaim locks after the server reboots. This option also controls the NFSv4 lease interval.
The default value is 90 seconds.
Specify the listener backlog,
listen_min_backlog, which is the number of connect requests that are queued and waiting to be processed before new
connect requests begin to get dropped.
Specify the number of seconds to wait before retransmitting a lock
request to the remote server. The default value is 5 seconds. Equivalent to the parameter in the NFS default
file.
nthreads Specify the maximum number of concurrent threads that the server can handle. This concurrency is achieved by up
to nthreads threads created as needed in the kernel. nthreads should be based on the load expected on this
server. If nthreads is not specified, the maximum number of concurrent threads will default to 20. Equivalent to
the parameter in the NFS default file.
Notes
A directory service that provides service name data base support must have the following service entries in its database:
Kernel Lock Manager logging can be enabled using By default, the variable is 0 which disables the Kernel Lock Manager logging feature. The
maximum value is 9. The value determines how detailed the log messages are.
AUTHOR
was developed by Sun Microsystems, Inc. and HP.
SEE ALSO
clear_locks(1M), kctune(1M), setoncenv(1M), statd(1M), fcntl(2), lockf(2), signal(2), getservent(3N), nfs(4), nsswitch.conf(4), ser-
vices(4).
lockd(1M)