Monthly Archives: February 2006
I made another mail configuraiton change last night on my home mail server. I did have NoSpamToday setup to forward all valid mail to 602LS on normal port 25. I decided to change this to a custom port so that I could setup port forwarding for the new port and use it externally from work or while traveling. This way, I can setup my mail client to use the new custom SMTP port on 602LS and all mail sent to my wife or users on my home network won’t have to get routed out to the internet, through rollernet and then back in to NST and then to their mailboxes.
SRS which can re-write the address headers and work around the fact that the messages were received by a forwarding mail server. I have yet to impliment a solution for this, as my mail servers use three SMTP servers and a workaround for all three is not yet available. Rollernet is working on implimenting SRS functionality on their servers, so I’m holding out for now hoping that when they get SRS working on their end, it will fix my issues of false positives on the SPAM filter. I think this will work in my case as NoSpamToday is my only SMTP server that does an SPF record check. So once its received by NoSpamToday, it should be forwarded to 602LS and accepted as valid mail. I’ll make another post with the results of this configuration once Rollernet releases this functionality.]]>
Liz and I have been spending a lot of time lately working on designing a website for her MOPS group. I havn’t had this much fun working on a site in a long time. We are using Mambo as the back end and tweaking the templates and HTML/CSS to suit our needs for the site. Its going to be really cool once its finished. Liz has put her creative edge into the site, and I’ve been doing all the code modifications. I wish I could do web design for a living, its a lot of fun, but I don’t possess the skill level with it to be able to make a career out of it. No wonder there are so many webmaster!
Liz and I were in the living room the other night working on a website together. Abby was in her room (supposedly taking her nap), and Sarah was in another room. Liz and I started to hear screaming that sounded like it was coming from outside the house. Like some kids were playing around and screaming outside. We didn’t think too much of it, since where we live there is lots of noise. The screaming kept going on and on and on, and Liz finally got up to try to see where it was coming from. A few minutes later, Abby comes crawling out of her room on her hands and knees sobbing and crying and trying to tell us something. Naturally we felt bad that she was crying but once we found out what happend we felt really bad. Turns out she had been playing in her room rather than sleeping. She had gotten herself stuck in her toybox which is nestled in her closet. The lid had closed on her and she had trouble getting out. She had been in there for quite a while screaming for Daddy and for someone to help her get out. When she came, she had red blots around her eyes where it looked like blood vessels had burst under her skin. Obviously she was terrified and was screaming for help for a while. The screaming we had heard was Abby, calling for help. But due to the location of her toybox and the loud area in which we live, we didn’t even consider it could have been one of our kids in trouble. We snuggled her for a long time trying to calm her down. It didn’t take her too long to get over it and get back to her old self. She started laughing about her “hinney” sticking out of the toybox while she was trying to escape. She and Sarah thought this was very funny once she calmed down. Once she came out of her room, Abby would try to tell Sarah what happened and start crying again, so it obviously upset her. Poor thing, we felt so bad about this. I especially felt bad when she said she was calling for Daddy. My child was in “trouble” whether or not we think its that severe, and was counting on me to help her out. I never showed up and hate the tought of disappointing her. From now on, there will be no more playing in the toy box!]]>
I was up until 1am this morning working on a web design project that Liz got me into. She would be working on th creative visual appeal of the site, and I’d be tweaking the HTML, PHP and CSS code to make the changes she wanted. It was the most fun I’ve had working on a site in my life. I’m developing a few sites using Mambo and Joomla which is pretty cool stuff. I like it better than PHP nuke or Post nuke or any of those.
Microsoft Active Directory:
My home network is built on Microsoft’s Active Directory. I use active directory to organize my user accounts (all two of them), my computer and group policies. With group policies I can set common variables for all my workstations, servers, etc. This way I don’t have to hand configure everything, its all automatic. Group Policies are a great way to manage your network workstations or servers. There are other solutions here, some people like to run Linux at home, and I’ll admit, I do too from time to time. I love linux, but there are still too many apps I use that require Windows. From time to time I demo some of the latest Linux distributions and try things out. I think its great, and if I had a 4th computer to run it on, I’d probably run a linux server or desktop as well. Some people like novell, some people like MAC, its up to you. This is just how I am doing thing. I have group policies set to add customization to my desktop mainly. Things like a browser title, automatic update settings, common software distribution, etc.
Domains, e-mail and more:
I guess I can’t go much further without explaining how I also do my domain names and websites. I’ll write more about this topic later on as a how to and what you should know for getting your own domain and website. But for now, I’ll keep it simple. I own several domain names which I use for various purposes. I have one domain that is for all my server equipment, like my hosting server that hosts my website and some other websites I host for people (for free unfortunately). These servers are in a data center and I simply “rent” the server from them on a month to month basis because its cheap and does what I want it to do. Plus they take care of maintenance and problems. Then I have a primary domain name I used to use for my hosting company’s website. The backend server domain ended with a .net and the primary domain is a .com. These extensions can be anything you like, but I stuck with a traditional format. Then I have a third domain for my personal website which is mainly for my family and my blog, etc. Here is where the bulk of my incoming and ougoing e-mail comes from, the other two domains are mainly for servers and a now closed hosting company. I do have some other domains, but don’t really used them yet. I’ll be expanding that later on as well.
So now you know I have a shared hosting server which hosts my websites and most functions of my domain names. Now when it comes to e-mail, you’d naturally assume this server also handled mail for my domains as well right? If you said yes, you’d be wrong. I’m using a service called Rollernet which is a mail forwarding service. Since my ISP restricts incoming traffic on port 25, it was necessary to setup SMTP on a non-custom port. However, this causes a problem because when someone on the internet sends me an e-mail, most mail servers only send mail on port 25. So if I’m running SMTP on a non-custom port, how do I get my mail? Here is how. Rollernet’s servers are listed as the MX records for my domains. This means, that when you send me an e-mail, its actually received on port 25 by rollernet. They take the mail, queue it, do some scans on it for viruses, spam etc, then they forward that mail to my home mail server on a custom SMTP port. Of course I have this port setup in my cable modem and firewall to allow it to be forwarded to my mail server which resided on my LAN. Now here is the complicated part. My home mail server received mail on a custom SMTP port and is received by NoSpamToday, which is my SMTP level SPAM filter. NoSpamToday (NST for short), filters for SPAM, viruses etc, and basically makes sure that the message is valid before it allows it in to my mailbox. Now NST is not a mail server, its just a SMTP server, so another component is needed here, thats where 602 Lan Suite (LS for short) comes in. NST received a message for me on a custom SMTP port. Once it makes sure that the message is valid, it then forwards that message to 602LS which receives the message on the standard SMTP Port 25. 602LS receives the message and performes a few checks of its own, like scanning it again for viruses, doing aother SPAM check and finally delivering it to my mailbox. 602LS also has a built in webmail server, so I can check my webmail from anywhere in the world. This is also where port forwarding comes in as the ports for webmail need to be setup to route to my home mail server from the outsite. Using my public DNS zone, I can add a record for webmail to my domain, so I can go to http://webamil.mydomain.com/mail and get to my web interface. This way I don’t have to use DynDNS or any of those services, since my public IP on my cable modem rarely changes. Now if it were to change, I’d have to manually update that in my DNS zone. So watch out for that if your using this scenario. I am aware of it and know what to do, so for me its not a big deal, but if your new to this, don’t set this up and wonder why it breaks 9 months later. Keep an eye on your public IP.
Lets now talk about outgoing mail. I don’t know if your like me, but I find myself in situations at work and abroad where I find that my company network or hotel network restricts SMTP servers to their own servers and won’t let you send mail using your own SMTP configuration. For example, at work I run a simple server monitor that sends alerts. But my company has a firewall in place that limits outgoing SMTP traffic on port 25. Now I bet your wondering where the SMTP component from IIS comes in to the picture from my previous post. Here it is. I am running IIS on my mail server but only the SMPT component. So I setup Microsoft’s SMTP service to listen on a custom port (different from my incoming SMTP port for normal e-mail from Rollernet). This way, I can setup my monitoring server to use my custom SMTP server at home to send the alerts. So in my situation, my monitor program detects a problem with a server in my office, it sends an alert to my home mail server on a custom SMTP port. My SMTP server then relays that message to my shared hosting server which then sends it to the desired recipient on a standard SMTP port. This way, I can use SMTP wherever I am, still get my messages or alerts sent and accomplish my tasks. This custom SMTP service is protected by a username and password and relaying with it is denied. Relaying on NST is also forbidden. Ok, so how about my home PC? Ok, simple, we use outlook on our home PC, so outlook is setup to send/receive mail from 602LS through POP3 and standard SMTP. We send a message from outlook, it is received by my home mail server on port 25, which then forwards that mail to my shared hosting server. Some ISPs also restrict outgoing SMTP traffic, so here you may need to setup a custom port on your public SMTP server and configure your mail server to send all outgoing mail over a “SmartHost” or custom SMTP configuration. My shared hosting server then delivers the mail over standard SMTP to the recipient’s mail server.
So in summary, yes this is a complicated setup, and no it may not be for everyone. But I will say this, there is a degree of pride that goes into setting soemthing like this up. Now I’m a Microsoft Engineer, so I’ve been doing networking for a long time. No this is not the way to go about setting up a business or large company. Obviously I’d recommend using Exchange or more powerful mail servers and betters ISP connections. But if your a techie and want to setup a really cool home network, this guide might just help point you in the right direction.
Lets talk remote access. So how do I manage this home network when I’m not home. Easy, RDP. There are lots of people around that don’t like RDP, its not very secure, and has its issues like any other software or technology. For me however, its perfect. I simply forward port 3389 from my cable modem to my firewall and from my firewall to my PC, I can remotely manage any machine on my home network. Now I took it a step further, and actually setup a custom RDP port on my other machines, like my servers and second desktop. This has the advantage of being easy to individually RDP into any machine on my home network without first having to remote into my home pc and then into another machine. In conjunction with DNS for easy naming, its a snap. All you need to remember is the custom port number for each machine. I only have a few so its no big deal, if you have many machines I’d recommend finding a better way, such as VPN. Through RDP I can remote control, and virtually manage any server or desktop on my home network.
Web management: I also use a program called Remotely Anywhere (www.remotelyanywhere.com). Its a great application that runs as a service on Windows. With it, you can remote control, Transfer files, totally manage all aspects of the machine right from a web browser. Its very robust and powerful, with tons of additional features too numberous to mention. Its one of the best web based remote control/management solutions I know of. This can also be setup on a custom port, so it will need port forwarding configured for it as well.
FTP: I used to have a NAS server with FTP setup so I could FTP directly to my RAID5 storage device. Now that its gone, I don’t really use FTP anymore so I removed it. I use an FTP site on my shared hosting server temporarily if I ever need to send anything through FTP. I can grab it from home later.
Internet Access: Because my cable modem and firewall do NAT, its very easy to provide for internet access to my workstations and servers on my home network. The firewall is the gateway on my network, and Microsof’t DNS handles all DNS related operations on my network. My DNS server is configured to forward all requests for external host names to my ISP’s DNS server. It then caches the results and can reply much faster to any requests my workstations or servers make. Internet access is basically a simple NAT solution provided by my firewall and cable modem.
Points of Failure:
With a system like this there are other considerations that need to be taken into account. Amoung them are power, redundancy, damage, replacement, etc. For example, if my power goes out what happens. Well for me I have my critical equipment on a UPS. Since this is a home network and not a critical system, the UPS will keep my servers and internet connection up and running for 5 minutes. This should be sufficient as long as the power isn’t out for long, which is isn’t usually. What if my firewall or cable modem goes bad. Well then I have a problem, as with my ISP I have to have them come and activate a new cable modem. So I’d first have to buy a replacement and then have them install it. This can be done usually by the next day. So what if my mail server or other network equipment is damaged. Well, for mail, if my home mail server becomes unavailable, mail will queue at rollernet, so I won’t loose any e-mail. I can even redirect that mail to my shared hosting server if I wanted to so I could get to it. If some of my network gear fails, it will obviously need to be replaced. I’d try to repalce it with exactly the same modem so that if it had a configuration with it, I could easily restore a backup config file to immediately get my network back up and running.
Security: What about security, how secure is this setup? Very secure. Even considering I have ports forwarded into my LAN from the outside. This often makes security experts very nervous and for good reason, but again, this is not the NSA, I don’t have anything on my home network worth anything to anyone but me. That is not an excuse for having bad security. First, I have a double NAT solution, so even if someone could hack in past my cable modem, they couldn’t get further than my firewall. If they could get past my firewall by some miracle, they would not be able to access anything on my network, since all network traffic between workstations and server is encrypted through Kerberos. The worst they could do if map out my network and find my IP addresses. DOS attacks are also a possability, but there isn’t much that can be done about that anyway. Again, I’m not saying good security isn’t important, and the measures I’ve taken are sufficient for my needs. Please don’t think I’m advocating bad security measures.
Thanks for taking time to read this post, I know it was long. Keep an eye out for more tech posts in the near future. I’ll also post some images giving you a visual of how all this works. Here is a simple visual aid of what I’m talking about above.]]>
Older model Compaq IPAQ desktops (small form factor, single drive,
Microsoft Windows Server 2003 Standard Edition (NOT FREE)
NoSpamToday – http://www.nospamtoday.com (FREE for 10 email addresses)
Microsoft IIS (only the SMTP component) (FREE if you own OS)
602 Lan Suite – http://www.602software.com (Free for 5 mailboxes)
AVG antivirus (for scanning incoming e-mail and server drives) – http://www.grisoft.com (Free)
Remotely Anywhere (not free)
Single 2003 Active Directory domain
Using Microsoft DHCP and DNS
Custom built home computer, 3.2GHz CPU, 1GB of OCZ Platinum DDR-800 RAM, approximately 800GB of onboard storage (sata 150) running Windows XP professional with SP2 as a domain member.
I used to have a Dell PowerVault 705N NAS server. It had 300GB of RAID5 storage but I very recently sold it on eBay.
Network (IP) Services used:
RDP (3389) (remote desktop)
Remotely Anywhere (custom)
Ok, this is the nuts and bolts which make up the core of my home network. In the next post, I’ll discuss how it all works, and ties together.
1. First, lets talk basics. Connection type for example. I’m using a cable modem connection to the internet at home from Bright House Networks (RoadRunner). I only have a 5MB down and typically 45kbs up. I’m planning on upgrading this soon to 10MB down and 1MB up. Speed here is important, because getting into your home network is fast, incoming speed (download speed) is the fastest. So if your sending your home server a file or receiving a large e-mail attachment, your going to need that faster bandwidth. Upload speed is equally important, because any mail you send out of your home network or file your uploading are going to be slower due to upload speed restrictions from your ISP. I highly recommend getting the faster upload speeds if your going to attempt anything remotely similar to what I’m about to explain to you.
Cable Modem – Motorola SBG900 (previously used SBG1000)
Firewall/router – Netgear Prosafe firewall/router (not wireless)
Wireless AP – Intel 802.11G Access Point
Dell 16 port switch
The cable modem is your gateway to the internet. If your like me, your cable modem uses NAT (Network Address Translation), and has a built in switch. This is useful because you can directly connect your various devices to your cable modem if you wish and not need to purchase a standalone router to share your internet connection. Note: your standard ISP cable modem probably won’t have this feature. Also, if you use the USB cable to connect to the internet, none of what I’m about to write about will work for you. You must go ethernet if this is to work. Plus, USP doesn’t use NAT (typically), you will end up getting the public IP of your cable modem if you go that route. This will open up your PC to attack from the outside and is not secure.
What I do:
I have a variation of a DMZ setup on my home network. My cable modem has a switch so I can use it to connect any devices I don’t care about and easily want to make accessible to the internet. One of those ports (I have 4), goes into the internet interface of my netgear firewall/router. So to the firewall, my cable modem is the gateway to the internet. I let DHCP give the firewall/router’s public (internet) interface its IP address. You can set this to static if you want, but if you get a firmware upgrade or your cable modem gets an update, your likely going to loose any port forwarding entries or custom setup, so using DHCP will save you time later, and keep your home internet connection from breaking. (I’ll tell you a story about this later). Now the firewall/router also has a LAN ethernet interface, so I have a cable (cat5 ethernet), going to a 16 port switch for other devices to connect to. In my cable modem, I setup all the outside services I want to have available to forward those ports to my firewall/routers internet interface. This is still a private address (non-routable), but will still work as intended. So the cable modem is listening on various ports for various services I have running on my home network accessible via my cable modem’s public IP address. When the cable modem received traffic on a particular port, it has a port forwarding entry that says, “ok, you ware coming in on this port, you go to the firewall”. The firewall is sitting there listening for those same ports, and has various rules setup to deal with traffic on different ports. So when traffic comes from the cable modem on a particular port, the firewall says “Ok, your coming from someone I trust (the cable modem), on a port I know about, and this port is supposed to go to this IP address on the LAN. The firewall then forwards that traffic to the server/device or PC on my LAN that I want it to go to. Port forwarding is key here, so first you setup port forwarders in your cable modem to forward to your firewall. The firewall then needs to have rules setup for the ports you want to use, specifying which internal (LAN) host you want the ports forwarded to. Examples of ports to forward are POP3 for e-mail, SMTP (although usually must be on a non-standard port), etc.
Now for wireless connectivity, my cable modem does support 802.11G wireless access, but it would be access to my DMZ, which is not helpful to me because the things I want to access when using my wireless are on my LAN. Now I could setup more security and custom routes, to make this work, but its much easier to just throw an 802.11G access point into the mix and set it up on your LAN. This way you get access to only the network you want access to. I also don’t typically encrypt my traffic using wireless only because I’m not doing anything secure or sending authentication or password traffic in the mix. Typically its good to add encryption to your wireless traffic. I use MAC address filtering to allow only wireless devices I know about access to my LAN. This in itself can be forged and hacked, so be careful when setting up security on your wireless network. Don’t just go buy a wireless Access Point and throw it in your LAN without configuring it. It will grant access to anyone with a wireless device to your network resources.
Ok, so thats the nuts and bolts of what I’m using for hardware and a touch of networking on my home network. Next we’ll talk about server, software and services.]]>