Re: [Leaf-user] portfw to *multiple* hosts ???
On Sat, 29 Dec 2001, Scott C. Best wrote: Paul: Heya. Notso left field, really. I've used ipfwd to forward IPSec packets (protocol 50 and 51) to my NAT'd LAN's broadcast address...and IPSec clients on that LAN can handle it. Of course...IPSec has some sense of state of a connection beyond just the packets. Webservers don't. I wonder what would happen with some game servers, or VNC. Hmmm... Webservers have a very clear sense of 'state' in their connections beyond just the packets. They use TCP, which maintains a pipe long enough to support a request and reply between the client and the server. The statelessness attributed to webservers has to do with the fact that these connections are generally very short-lived... under http 1.0, only a single request/reply is supported before the connection is dropped. Multicasting is useful for now data... radio broadcasts, for example, where if you don't get the sound from a few seconds ago, oh well, because you won't be able to catch up to the sound corresponding to now if you try to listen to re-transmissions of packets sent then. Large file transfers cannot afford to omit retransmissions, and some amount of delay is worth the delay involved in making sure the file gets through intact. One level of difficulty is variable reliability of transmissions. Keeping track of the fact that after transmitting 300 packets, destination A needs a retransmission of packet 293, but destination B needs packet 297, and destination C got it all so far, is outside the scope of normal TCP connections. They only work on a one-to-one basis. Another level of difficulty is the fact that the application level protocols are typically designed for point-to-point communication... the communication initiator does not expect to hear several voices answer when it says Hello. Therefore I think Charles' suggestion to have an application-level repeater would be the only way to construct this data flow. -Scott Maybe this is way out in left field, but could this be a potential application for multicasting? paul ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user --- Jeff NewmillerThe . . Go Live... DCN:[EMAIL PROTECTED]Basics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...2k --- ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
??? Please explain a bit more about exactly what you're trying to accomplish... Large medical images -- some approaching gigabyte sizes. The internal network connects multiple facilities. The images may need to be shared across multiple facilities. Our preferred solution is to put one (1) copy of each image on a large and robust fileserver inside their network. The catch is, they are using proprietary systems for viewing and analyzing the images and we may not be granted access nor information adequate to implementing our preferred solution. Currently, the remote sources are using their proprietary systems (black boxes) to auto-magically transfer the files directly to one (1) proprietary system inside our customer's network. Yes, this looks everyway like ftp -- except the proprietary system vendor says, no, it is not that simple ; When one of these images is needed on another proprietary system inside this network, somebody needs to push the required file to another proprietary system. Our customer wants ``pull'' access from any given system. In brainstorming alternatives, this occured to me: send images | V internet | V firewall | - | | | V V V host_1host_2host_n ... Regardless, whether or not this is the best solution for this application, how can this be done? Well, it sounds like you're in black-box hell : My current understanding of your problem (please correct me if I'm wrong): - A remote, black-box system pushes images to one black-box system in your customer's network (let's call it Master) - Your customer can push images from the Master black box system to several other black-box systems (Slaves) - Your customer wants to be able to view images from any slave system w/o having to push the file from the master system (ie pull, not push) Without an understanding of the black-boxes involved, there's not a lot you can do here. If I am now understanding your initial question properly, you are asking if it's possible to somehow get the remote system to push the image content to multiple systems on your internal network. The simple answer is no...there is no straight-forward network slight-of-hand that will allow you to trick one system into talking multiple remote systems at the same time. The more complex answer is maybe, and depends a lot on unknown details of your black-box systems... I can think of a few things that could potentially help you: 1) Get the remote system to push the data to all clients. This will chew up lots of internet bandwidth, and will either require multiple external IP's, or control over which ports are used to connect (assuming you're running a masqueraded internal network). 2) Get the Master internal system to push the data to all slave systems. This may be possible to automate, or it may require a console jocky...depends on unknown (to me) details of your black-box systems. 3) Whip up a Black-Box emulator, that can talk the file-transfer protocol used by your systems. Have your remote system transfer files to your fake system, then push the data from there to all internal systems, as required. This will require network programming (not too tough in modern languages like Java, Perl, Python, c) and an understanding (or reverse engineering) of the file transfer protocols used by your equipment. 4) Read through the docs for your black-boxes, and see if they support any kind of image-server access. I think this is really what you want...a central image server. If you're lucky, you can do this with a *nix box. If you're unlucky, you'll need a propriatery box from your vendor for the other systems. If you're really-really unlucky, there's no support for any kind of 'pull' technology on the individual Slave stations. NOTE that everything but the last option requires lots of storage capacity on each image station, since each station is storing copies of all the image files. Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
Charles Steinkuehler wrote: ??? Please explain a bit more about exactly what you're trying to accomplish... Large medical images -- some approaching gigabyte sizes. The internal network connects multiple facilities. The images may need to be shared across multiple facilities. Our preferred solution is to put one (1) copy of each image on a large and robust fileserver inside their network. The catch is, they are using proprietary systems for viewing and analyzing the images and we may not be granted access nor information adequate to implementing our preferred solution. Currently, the remote sources are using their proprietary systems (black boxes) to auto-magically transfer the files directly to one (1) proprietary system inside our customer's network. Yes, this looks everyway like ftp -- except the proprietary system vendor says, no, it is not that simple ; When one of these images is needed on another proprietary system inside this network, somebody needs to push the required file to another proprietary system. Our customer wants ``pull'' access from any given system. In brainstorming alternatives, this occured to me: send images | V internet | V firewall | - | | | V V V host_1host_2host_n ... Regardless, whether or not this is the best solution for this application, how can this be done? Well, it sounds like you're in black-box hell : Do I detect a bit of empathy from somebody who's been here before? My current understanding of your problem (please correct me if I'm wrong): - A remote, black-box system pushes images to one black-box system in your customer's network (let's call it Master) - Your customer can push images from the Master black box system to several other black-box systems (Slaves) - Your customer wants to be able to view images from any slave system w/o having to push the file from the master system (ie pull, not push) Yes -- this pretty much sums it up ; Without an understanding of the black-boxes involved, there's not a lot you can do here. If I am now understanding your initial question properly, you are asking if it's possible to somehow get the remote system to push the image content to multiple systems on your internal network. The simple answer is no...there is no straight-forward network slight-of-hand that will allow you to trick one system into talking multiple remote systems at the same time. The more complex answer is maybe, and depends a lot on unknown details of your black-box systems... Yes, that is core to my question: can port forwarding process do a ``tee'' to multiple addresses? I can think of a few things that could potentially help you: 1) Get the remote system to push the data to all clients. This will chew up lots of internet bandwidth, and will either require multiple external IP's, or control over which ports are used to connect (assuming you're running a masqueraded internal network). 2) Get the Master internal system to push the data to all slave systems. This may be possible to automate, or it may require a console jocky...depends on unknown (to me) details of your black-box systems. 3) Whip up a Black-Box emulator, that can talk the file-transfer protocol used by your systems. Have your remote system transfer files to your fake system, then push the data from there to all internal systems, as required. This will require network programming (not too tough in modern languages like Java, Perl, Python, c) and an understanding (or reverse engineering) of the file transfer protocols used by your equipment. 4) Read through the docs for your black-boxes, and see if they support any kind of image-server access. I think this is really what you want...a central image server. If you're lucky, you can do this with a *nix box. If you're unlucky, you'll need a propriatery box from your vendor for the other systems. If you're really-really unlucky, there's no support for any kind of 'pull' technology on the individual Slave stations. NOTE that everything but the last option requires lots of storage capacity on each image station, since each station is storing copies of all the image files. Yes, these are the alternatives we are considering. Yes, all of these depend -- more or less -- on access to proprietary information to which we may or may not be privy ; Of course, the image/fileserver approach makes most sense from many perspectives. Nevertheless, the port forwarding ``tee'' feasibility is an interesting question, regardless of our current customer's predicament! man ipmasqadm, on my potato box, contains and interesting example, which may or may not shed light on this; but, which I also do *not* fully understand: ``Redirect all web traffic to internals
RE: [Leaf-user] portfw to *multiple* hosts ???
Maybe this is way out in left field, but could this be a potential application for multicasting? paul Paul M. Wright, Jr. McKay Technologies making technology play nice ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
Yes, these are the alternatives we are considering. Yes, all of these depend -- more or less -- on access to proprietary information to which we may or may not be privy ; Of course, the image/fileserver approach makes most sense from many perspectives. Nevertheless, the port forwarding ``tee'' feasibility is an interesting question, regardless of our current customer's predicament! But saddly, this is not generally possible. Certianly not without some custom code. Perhaps not at all depending on the details of the communication protocol (or at least not without a full application level proxy). man ipmasqadm, on my potato box, contains and interesting example, which may or may not shed light on this; but, which I also do *not* fully understand: ``Redirect all web traffic to internals hostA and hostB, where hostB will serve 2 times hostA connections. Forward rules already masq internal hosts to outside (typical). ipchains -I input -p tcp -y -d yours.com/32 80 -m 1 ipmasqadm mfw -I -m 1 -r hostA 80 -p 10 ipmasqadm mfw -I -m 1 -r hostB 80 -p 20'' What is this really doing? If I'm understanding this properly, it's doing load-balancing. This divides individual inbound web connections between two internal web servers. You want something that takes a single inbound connection (from the remote system to your internal Master system) and Tee's the connection to multiple systems, which is just not how TCP communications work... Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
Paul: Heya. Notso left field, really. I've used ipfwd to forward IPSec packets (protocol 50 and 51) to my NAT'd LAN's broadcast address...and IPSec clients on that LAN can handle it. Of course...IPSec has some sense of state of a connection beyond just the packets. Webservers don't. I wonder what would happen with some game servers, or VNC. Hmmm... -Scott Maybe this is way out in left field, but could this be a potential application for multicasting? paul ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
[Leaf-user] portfw to *multiple* hosts ???
Quite simply, what is the simplest, secure way to forward to two (2) hosts? There are probably better ways to accomplish the end goal; but, we have an application whereby we may need to push very large files from the internet to two (or, more) locations behind a Dachstein firewall. What do you think? -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
On Thu, 27 Dec 2001, Michael D. Schleif wrote: Quite simply, what is the simplest, secure way to forward to two (2) hosts? There are probably better ways to accomplish the end goal; but, we have an application whereby we may need to push very large files from the internet to two (or, more) locations behind a Dachstein firewall. What do you think? scp or https/PUT to separate ports (22 and 2022, or 443 and 4443, for example), one port for each host. The hosts could each see input on the nominal port (22 or 443). --- Jeff NewmillerThe . . Go Live... DCN:[EMAIL PROTECTED]Basics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...2k --- ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
Quite simply, what is the simplest, secure way to forward to two (2) hosts? There are probably better ways to accomplish the end goal; but, we have an application whereby we may need to push very large files from the internet to two (or, more) locations behind a Dachstein firewall. Simplest is to simply create multiple INTERN_SERVERS entries. You'll need to port-forward from different ports (or have more than one external IP address). Security of a port-forwarded service is only as secure as the service itself. If you really need tight security, you might consider an application level proxy. For instance, if you're port-forwarding to a MS IIS web-server, you might want to run all requests through a *nix based proxy that filters out (and logs) any *default.ida web-requests. If you can afford the overhead (and potential cost...there are many good application level proxies that are for sale commercial products), this can be a good way to shield yourself from various attacks, both known and (at least some) unknown...stuff like buffer overflow attacks, broken protocol attacks, and the like. If you want to use a proxy, just port-forward the service to the proxy instead of your 'real' server, and configure the proxy to talk to your real server(s)... If you're looking at pushing/syncing a large number of files to various remote sites, you might also want to look into rsync and/or ssh. If the files aren't terribly sensitive (ie can traverse the internet unencrypted), you can setup an rsync server at the 'master' site and sync all the clients periodically. This is more of a 'pull' architecture, but it can be made into a 'push' system by having the master run the rsync download command on the clients via ssh. If you need to encrypt the transfers, you can tunnel the entire session through ssh. You can keep security as tight as you want with proper ssh configuration...for something like this I usually disable general logins, setup ssh authentication by RSA/DSA keys only, and have the ssh session automatically invoke the proper behavior on the client end (ie fire off an rsync session in your case). This way, even if the master server is compromised, you won't autmoatically get user-level access to the clients...all you'll be able to do is force them to rsync to the master server whenever you want... You can also get rsync/ssh for windows (see cygwin http://www.redhat.com/download/cygwin.html ) if your network is of the M$ persuasion... Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
Quite simply, what is the simplest, secure way to forward to two (2) hosts? There are probably better ways to accomplish the end goal; but, we have an application whereby we may need to push very large files from the internet to two (or, more) locations behind a Dachstein firewall. What do you think? scp or https/PUT to separate ports (22 and 2022, or 443 and 4443, for example), one port for each host. The hosts could each see input on the nominal port (22 or 443). Yes, I see this; but, is there someway to accomplish this -- simultaneously -- with one (1) remote operation? ??? Please explain a bit more about exactly what you're trying to accomplish... Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
Charles Steinkuehler wrote: Quite simply, what is the simplest, secure way to forward to two (2) hosts? There are probably better ways to accomplish the end goal; but, we have an application whereby we may need to push very large files from the internet to two (or, more) locations behind a Dachstein firewall. What do you think? scp or https/PUT to separate ports (22 and 2022, or 443 and 4443, for example), one port for each host. The hosts could each see input on the nominal port (22 or 443). Yes, I see this; but, is there someway to accomplish this -- simultaneously -- with one (1) remote operation? ??? Please explain a bit more about exactly what you're trying to accomplish... Large medical images -- some approaching gigabyte sizes. The internal network connects multiple facilities. The images may need to be shared across multiple facilities. Our preferred solution is to put one (1) copy of each image on a large and robust fileserver inside their network. The catch is, they are using proprietary systems for viewing and analyzing the images and we may not be granted access nor information adequate to implementing our preferred solution. Currently, the remote sources are using their proprietary systems (black boxes) to auto-magically transfer the files directly to one (1) proprietary system inside our customer's network. Yes, this looks everyway like ftp -- except the proprietary system vendor says, no, it is not that simple ; When one of these images is needed on another proprietary system inside this network, somebody needs to push the required file to another proprietary system. Our customer wants ``pull'' access from any given system. In brainstorming alternatives, this occured to me: send images | V internet | V firewall | - | | | V V V host_1host_2host_n ... Regardless, whether or not this is the best solution for this application, how can this be done? What do you think? -- Best Regards, mds mds resource 888.250.3987 Dare to fix things before they break . . . Our capacity for understanding is inversely proportional to how much we think we know. The more I know, the more I know I don't know . . . ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user
Re: [Leaf-user] portfw to *multiple* hosts ???
On 12/27/01 at 10:21 PM, Michael D. Schleif [EMAIL PROTECTED] wrote: Large medical images -- some approaching gigabyte sizes. The internal network connects multiple facilities. The images may need to be shared across multiple facilities. Our preferred solution is to put one (1) copy of each image on a large and robust fileserver inside their network. The catch is, they are using proprietary systems for viewing and analyzing the images and we may not be granted access nor information adequate to implementing our preferred solution. Currently, the remote sources are using their proprietary systems (black boxes) to auto-magically transfer the files directly to one (1) proprietary system inside our customer's network. Yes, this looks everyway like ftp -- except the proprietary system vendor says, no, it is not that simple ; When one of these images is needed on another proprietary system inside this network, somebody needs to push the required file to another proprietary system. Our customer wants ``pull'' access from any given system. In brainstorming alternatives, this occured to me: send images | V internet | V firewall | - | | | V V V host_1host_2host_n ... Regardless, whether or not this is the best solution for this application, how can this be done? What do you think? This sounds to me like a case for rsync + ssh There is, if you need it, an rsync.lrp already - and of course, ssh.lrp. You could set up rsync either as a push or a pull alternative. As a case study, consider that there are many publicly accessibly rsync servers (the Linux kernel site kernel.org comes to mind...) If you could set up host_1, host_2, etc. to be rsync recipients, why not tunnel rsync via ssh through the firewall? -- David Douthitt UNIX Systems Administrator HP-UX, Unixware, Linux [EMAIL PROTECTED] ___ Leaf-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user