Simplysimple.info » DIY

Components of a Network

Benjy — Wed, 14 Sep 2011 08:11:00 +0000

The Network schematic showing how the parts of the network connect to each other are detailed in the next post. You might want to take a peek. In this post, I want you to get familiar with each of the devices and the connectors you would find on them and how they can be accessed for configuration.

Keep in mind that specifications and design change constantly. In addition, I will be mentioning only those features on the device that need mentioning, there may be other features which would be beyond the scope of this post. One more fact that needs mentioning is that the web administration interface on each device is different, so it is impossible to put up screen shots and provide customized guidance for each and every model out there, so treat these series of posts as a generic guide.

Switches

Let us start off with the most familiar device that would be found on any network -the network switch. For those who have lived longer, they would remember the predecessor of the the switch -the hub. Don’t bother about the difference between a switch and a hub, the switch does the same thing that the hub did, only it does it better -and faster. Switches come in two types -managed and unmanaged, which can probably be explained off as programmable (managed) or dumb. We will be dealing only with “dumb” unmanaged switches.

When buying a switch, always buy a switch that has a metal box which would also have an inbuilt power supply, rather than switches in plastic boxes with bulky unwieldy external power adapters. The metal body shields the electronics from stray RF fields.

: A 10/100 Network switch without its rack mounting brackets attached. A more expensive and faster switch would be a Gigabit switch ( 10/100/1000)

The switch in the photograph above can also be rack mounted. The rack mount brackets are basically two L shaped brackets which attach to the switch body with four set screws each. You can see a picture below. All you will find on this switch, (apart from all the Ethernet sockets) is a AC power socket which is identical to the socket used for powering your desktop or monitor.

: Two switches mounted in a standard 19″ rack.

If you run out of ports on one switch, you can daisy chain another switch by connecting the last free port on the existing switch to an empty port on the new switch. However for optimum performance and reliability, this must be kept to a minimum and used only when there are two groups of computers in physically different locations such as on two different floors of a building, this saves on the number of wires but machines connected to the second switch will not be able to connect at the rated speed of the switch to a server on the first switch. (This is not true for expensive managed/stackable switches) This is why, it makes sense on a busy network to go in for the fastest switch that your wallet allows. If you expect your network to expand, it is good to buy a switch with additional ports (normally 16, 24, 48 etc) the additional ports also help if one or more ports fail in operation.

Wireless Access points or AP’s

Wireless access points act like wireless switches. They need to be connected to a port on a network switch before they can be usable. They can usually be configured in two modes -as wireless AP’s or as wireless repeaters. In wireless repeater mode, it does not need to be connected to a wired network switch. It repeats or amplifies and rebroadcasts an existing wireless signal. It works well for extending the range of an existing wireless signal. As with network switches, if the wireless signal is repeated from repeater to repeater in succession (which is like daisy chaining switches) you end up with degraded performance.

: A wirelss AP for Small office/Home (SOHO) use

You can see the connections in the photograph above. A LAN (Ethernet) port for connecting to the wired network, a DC socket (yes it has a boxy external power adapter) and a hole to access the reset button which resets the AP to factory settings.

: An AP for business use having primary and secondary antennae

AP’s come in two speeds -54Mbps (older AP’s) and 108Mbps. The above AP has its power supply built into it instead of an external adapter. Wireless access points are programmable and come with their own web interface through which they can be setup. This means that you can open Internet Explorer or Firefox, type the AP’s IP address in the address bar, enter the default user name and password when prompted and configure the AP according to your preferences. More on this later in a separate post.

Modem/Router

Finally we move on to the Internet modem/router. Since our target audience is basically Home/SOHO, we will only be discussing DSL connections under this section. All the DSL modems I’ve come across from MTNL, BSNL, Airtel etc.. seem to have their IP addresses set to 192.168.1.1 By connecting to its web interface through a web browser, it can be configured in a similar fashion to that of the AP’s above. DSL modems come with a reset button too, but keep in mind that initiating a factory reset on one of these boxes would wipe your ISP settings and you might need to call up your ISP (Internet Service Provider) to send their representative to re-program the modem.

: A buch of assorted modems/routers

Please note that both Modem/Routers and Wireless AP’s have built in DHCP servers. Some manufacturers foolishly leave them on by default to ease initial connectivity/configuration. Unfortunately sometimes they get reset and start broadcasting which causes chaos on the network as only one DHCP server should run on the network. If this DHCP thingy sounds like Greek, then you need to re-visit the post on networking basics as this is important.

Some Internet modem-routers come with a wireless antenna, I haven’t found this very effective. Avoid buying it except for home use. Some routers also have a few additional ports as they have an integrated switch which saves users from buying a switch for home use. For an office with a busy network, these are more or less toys. So choose the right stuff for your network.

How it connects together

Benjy — Tue, 09 Aug 2011 14:26:09 +0000

Back to Networking Simplified

You now know enough about your networking basics to start putting things together. However don’t rush into it as this post would be incomplete without the upcoming posts on wiring your network so that you can setup your wiring once and forget it for a long time to come. Then there are also power considerations to ponder upon. Networking equipment is very sensitive and your switches, routers and modems do not need too much of a spike to damage them making them die a premature death or making them unreliable due to reduction in capacity or intermittent operation. They are also affected by RF (Radio frequency ) and other strong electromagnetic fields, so whenever you have an option of buying any of these devices in a metal enclosure instead of plastic, go for it.

The network that we will be looking at will have a few computers connected through a network switch. Unlike electrical wiring, you cannot splice two or three networking cables together to extend your network. The signals on the 4 twisted pairs of the Ethernet cable must be amplified and then split into additional channels by a network switch.

To this network we will be adding an Internet connection so that everyone can browse the Internet, a network printer (an unlikely purchase for a small network) and one or more wireless access points to serve the needs of devices that connect through wireless, like mobile phones, laptops and other WiFi enabled hand held devices.

A simple office network (Network printer is optional as it is pretty expensive)

Note that the Network switch is at the center of everything and it ties your network together. It’s predecessor was called a hub for obvious reasons. A wireless access point works as a wireless network switch. All service recipients and service providers, connect to the switch through which information interchange takes place.

The main network consists of the Desktops, Laptops and smart phones. All others only add value to the network by providing services, For example, the Network printer provides print services and the Modem/router provides a gateway for Internet access and the wireless access point extends the network and provides network access to devices wirelessly.

This network diagram is only indicative of the connections, you can add more than one network printer or wireless access point. If you run short of wired network ports, you can connect another switch to one of the free ports on the existing switch and extend the network. However I advise against this as it creates multiple points of failure on the network, so preferably try and use a single network switch for such a small network.

If you remember what a DHCP server was, [ Go back to Networking basics and brush up if you've already forgotten] then it needs to run on a server. All of the devices -the router, wireless access point, network printer are servers and they mostly run Linux from flash memory i.e they have no hard disks. So all of them have DHCP servers built into them. This means that you must configure and turn on only one DHCP server for your network or there will be chaos. The printer might be turned off if there is a paper jam. Your DHCP server must always be on so the ideal place to run the server would be on your Internet router. In addition, each of these boxes also run a web server -a complete “website” is running on each of them. This is its web administration interface. we can configure/program these devices using a web browser such as Firefox (preferred).

We will look closely at each of these devices in the next post and also on how to configure the devices.

Networking basics

Benjy — Sun, 07 Aug 2011 06:41:41 +0000

Back to Networking Simplified

Remember, you are going to get only as much basics as you need to set up your own Home or small office network. If you need additional information please look elsewhere as this post is written with a particular audience in mind and too much information is overwhelming and confusing.

I’m assuming you want to do one or more of the following.

Share your Internet connection on your network
Share printers
Share folders
Connect to your network via wireless on mobile devices

This is what most people would want to do. Any thing beyond this is also beyond the scope of these series of articles titled Networking simplified. However without a wee bit of theory, understanding the practical aspect will be a bit difficult, so here goes.

To start with, we all know that each of us needs to have a unique postal address or email address for either post or email to be successfully delivered to us. This is no different for computers, only in their case, these addresses are called IP addresses short for Internet Protocol addresses. This is standard for computers communicating over the Internet and we will use the same to setup our local network. These are not only required for delivering email, but for any kind of communication between two computers.

The format of an IP address is XXX.XXX.XXX.XXX . If any of the numbers are less than three digits then zero’s to the left of the number are omitted. There are various classes of IP addresses and we will not go into details. We will be using only one type of private addressing scheme often called the class C address.

For any class of address, an address can be assigned to a machine either manually (statically) or dynamically through a DHCP (Dynamic Host Configuration Protocol) server. Assigning computers static IP addresses is tedious and dumb and is only done for machines which need to be uniquely identified such as a server or other networking appliances providing services to other computers on the network.

An example of an IP address would be 121.245.8.55 This is the IP address of my computer right now and is assigned to me by my Internet Service provider (ISP) from their DHCP server. This changes every time (mostly) I connect to the Internet and when I disconnect, it is assigned to the next subscriber who connects to my ISP.

This is a common way of assigning a limited range of addresses. The IP addresses are recycled and leased to new users as and when they connect to the network. This way, if we have 100 laptops on a network, and only 40 of them on average are in office at any given time, I can safely setup a pool of 60 addresses on the DHCP server. In addition I wont have to keep a list of who has which IP address (for all 100 of them) which is crucial if IP addresses are assigned manually as two computers having the same IP address will have an IP address conflict and will lose their connectivity or work intermittently.

To connect any machine to the Internet, we need to specify (either statically or dynamically through a DHCP server) 4 different addresses. On a Linux server, you will need to setup two more, however we will list them and forget about them as we are not setting up such servers in this series of posts.

An IP address 192.168.1.11
The Subnet mask 255.255.255.0
The default gateway 192.168.1.1
One or more DNS servers 8.8.8.8

Like we said, on Linux machines with a static IP, we need to specify two more addresses

The Network address 192.168.1.0
The broadcast address 192.168.1.255

In our tutorial, we are going to setup the 192.168.1.0 network. I chose this because traditionally Airtel, MTNL and BSNL modems have been shipping with modems configured with a static IP address of 192.168.1.1. In case we change this address, there will be chaos and your network will break every time the service technician comes to repair any fault or to troubleshoot your modem.

In this network range, you can address up to 254 different machines. This is because the highest number 255 is reserved as the broadcast address and the lowest 0 is reserved as the network address. In addition it is good practice to reserve the first few IP addresses for devices requiring static IP addressing. I reserve the first 10 to be on the safe side. This leaves you with an effective IP address range from 192.168.1.11 to 192.168.11.254 – 244 addresses, more than sufficient for a small office network.

Coming back to the 4 essential addresses listed above, the first, the IP address is unique for each computer. If the first one ends with 11, the nest one ends with 12 and so on up to a maximum of 254.

You can forget about the subnet mask, the default gateway and the DNS server as you can safely use the given values and expect your network to work, so don’t break your head over it. For the above service providers, the gateway is always 192.168.1.1. The DNS server 8.8.8.8 is the address of Google’s public DNS server so it should work everywhere. So apart from the main IP address, all the other settings will be the same as shown above for all the computers on the network.

We will look at DHCP (automatic addressing) in more detail when we actually configure our network. This is enough theory to get started.

Remember that you can have a working office network with addresses 1 and 2 stated above, but you will not have Internet access, until the 3 and 4 addresses are in place. The gateway tells the client computers through which device it needs to send its external (Internet) requests to, while the DNS address is the address of the DNS (Domain Name Translation) server which translates Internet domain names such as www.eha-health.org to an IP address such as 207.58.183.47

What is a server?

Benjy — Sun, 07 Aug 2011 03:32:22 +0000

Back to Networking Simplified

This is something I get regularly quizzed about. When the server is installed in a networking cabinet and looks radically different from a desktop in its form factor, people have no questions about it. However when its form factor changes for example in the lower range servers and someone looks at a tower server, which more or less resembles the CPU (box) of a desktop computer, or if they know that a server operating system can be installed on a Desktop computer and it can then be used as a server, then a lot of questions arise in people’s minds. These questions often come into the minds of the bean counters or that of the top level management who are not well versed in this field of operation, so it becomes difficult for them to justify the cost of a server when presented with the fact that theoretically the same software will run on a desktop as well .

Just a note before we proceed… Remember that when we speak of servers and desktops, it can mean either the hardware or the software. When a server operating system is installed on a desktop, it performs the role of a server and is therefore a server although practically it is only a Desktop masquerading as a server.

A desktop operating system also has some server functions built into it. When a Windows, Mac or Linux desktop shares a printer with other clients on the network, it works as a print server, when it shares folders over the network it emulates (with limitations) a file server and when it shares its Internet connection with other computers it acts as a proxy server.

A server is a machine that is dedicated on the network or Internet to provide services to its designated clients. Its services depend on the configuration of its Operating system (e.g Windows 2008, Linux, Unix, etc…) and applications if any running on it.

So why not run from a Desktop?

For the sake of illustration, let us assume that a certain company runs its web site, email and some other critical services from a single server (Not a very good practice). This now becomes a single point of failure for the services required by that company. Although the same server operating system will install on a Desktop computer and run the required services, the reliability and performance of the services will be poor which would become a liability for the company. Let us look at some of the major differences between a Desktop machine and a server.

The server is designed to run 24 x 7 x 365. While it is possible that a Desktop can do that too, the desktop was never specifically designed for it.
A server and desktop are designed for different applications, so they perform as they are designed in those areas. Installing a Windows or Linux desktop operating system on a server and using it as a desktop will be a miserable experience with bad graphics performance. Similarly running a server on a Desktop will result in poor performance as that machine was not tuned in specific areas required for working as a server.
Because a server is designed to run 24 x7 x 365, the air flow passages, position of boards and fans are designed keeping that in mind. A desktop has no such design consideration as most users shut down after work. In fact some of the assembled desktops overheat even under normal operation.
A server supports a faster multi-core processor and oodles of RAM and has very reliable high speed network interfaces. It can usually read and write from disk much faster too.
A server boots slower and intensively tests its hardware before starting up. Our server takes close to 5 minutes to start up, but when it is up, it is blazing fast.

In addition to this, a server may have any or all of the following features.

More than one fan. In higher end servers, the fan speed and temperature is monitored. In addition, the fans can be hot swapped -pulled out and replaced while the server is running so that the server needn’t be turned off for fan replacement.
RPS -short for Redundant Power Supply. This means it has more than one power supply. Often these are hot swappable too, which means that you can unplug and pull out the damaged RPS unit and replace it with a spare with the server continuing to run from the other power supply.
Hot swappable hard disks usually using some kind of redundancy like hardware RAID. Without going in to too many details, RAID (or rather some levels of RAID) allow you to write to more than one disk at a time thus preventing data loss if one of the hard drives fail. The failed disk can be (if hot swappable) ejected out of its bay and then replaced with a spare after which the new RAID member (disk) can be rebuilt or synced to the state of the failed disk without data loss and without turning off the server for disk replacement.
Advanced diagnostics and fault alert systems and a remote console for administering the server from another machine on the network. This enables you to in addition to many other tasks to forcibly shut down or restart an unresponsive server without going through the operating system or having to go to the server room.

A server would usually be housed in an air conditioned cabin to keep it running within its optimal temperature range. This would also reduce the amount of dust which is as great an enemy as high temperature is.

In the next post in this series we will be looking at Networking basics.

How a computer works

Benjy — Fri, 05 Aug 2011 17:12:43 +0000

Back to Networking Simplified

We can divide the logical parts of a computer into five horizontal slices.

Hardware (Right at the bottom)
Firmware or BIOS software (Basic Input/Output System)
The Operating System
Driver Software ( Device drivers)
Application Software (Topmost)

The Hardware is obviously what you can see, feel and touch. We can sub divide this into two parts.

Hardware belonging to the computer without which it won’t run.
and Hardware peripherals that extend the use of the computer like scanners, printers, pen drives etc…

The firmware or BIOS software is software that is flashed electronically into the BIOS chip. When the computer is first turned on and electricity starts flowing through its components, the computer wakes up. The CPU or the Processor -the heart of the computer, in spite of having the ability to do things at tremendous speeds wakes up more or less like Mr. Bean -or for those of you who hate him, like a new born baby which only knows to suckle (apart from cry). The CPU too knows only to look up a certain address and execute the first available instruction there (more or less). This address falls in the range allocated to the BIOS software or rather the address belongs to the first instruction of the BIOS software.

Once the BIOS software is loaded and executed, the computer gains enough brain power to do simple things like a POST (Power On Self Test) a memory test and numerous other things. If any of these fail, before it attains the capability to inform the user in a better way -like printing it on screen, it beeps in a coded way to let the user know what part failed like RAM, Display etc… just like a baby cries due to its inability to use words for communication.

If all went well, it starts the boot process. It checks the boot order saved in the BIOS settings by the user and tries the boot devices (e.g, DVD drive, Hard drive, USB stick, Network) one by one in the set order until it can successfully boot from one of them. If none of them work, it terminates the process with a boot failure message. In case it succeeds, the BIOS program hands over control of the computer and its resources to the Operating System.

It doesn’t matter what your operating system is -Windows, the Mac, Linux, they all do the same job -of that of an operating system. They interact with the hardware on the users behalf and on behalf of the driver software and application software.

Driver software aka device drivers, describes the capabilities of hardware to the operating system and allows the operating system to optimally control the hardware. The hardware could be part of the computer or could be an external device like a printer, scanner, Bluetooth dongle, mobile phone -any extra hardware connected to the Computer. You should be aware that quite a lot of hardware drivers are integrated into the Operating System. Newer Operating Systems like Windows Vista and 7 try to connect to the Internet and try to get the driver software for you while Linux seems to recognize a large percentage of hardware on its own.

You can think of driver software like the air in air brakes. The Operating System here is the drivers foot on the brake pedal and the hardware peripheral is the brake drum on the wheel. If the air (driver software) is missing, the brake will not be able to be operated. It will remain a cosmetic add-on on the vehicle, exactly like your inoperative printer on your desk whose driver CD you lost.

Application software, anything you install on the Operating System like MS office, Open office, your antivirus software, web browser and of course software like Limewire and Bittorrent (I know you do that you crooks!)

Now that we have a general idea of how a computer works, we can proceed to the networking theory after a quick look at servers.

Networking Simplified

Benjy — Fri, 05 Aug 2011 17:01:15 +0000

I work for an organization which is an association of about 20 odd mission hospitals spread out over suburban and rural India. Although our hospitals are spread out more over North, Central and North Eastern India, there are other similar organizations who call me up for consultancy regularly who cover some of the parts left out by us.

As a charitable organization, our IT struggles are different from that of the cash rich corporates. Our solutions too cannot always be bought at a price as that price is usually beyond our reach or grossly unjustified for the kind of work that we do. Open source software has opened the door to us for mainstream technology, but to get enterprise level technology to work for us takes a lot of hacking and re-shaping to fit a product into our mold. In addition we struggle with issues like erratic power supplies -blackouts, brownouts, surges and spikes that not even industrial level PLC’s have been able to handle reliably.

This series of posts has come about by popular demand and I have striven to keep it as simple as possible for people with the aptitude, to be able to follow the instructions successfully. While reading these posts, don’t read the text and the descriptions too hard but soak in the basics as the absolutes keep changing with every product that is released while the underlying concepts change much more slowly. I know of IT managers who throw up their hands in despair if the replacement modem came with a firmware update which showed the same menu in green instead of blue. In short -don’t emulate them.

Finally stop short where you feel that the complexity is too much for you to handle as although we start off with easy things, we must tackle the complex issues sooner or later.

Most of the voltages we will encounter will be below 12 volts which makes it pretty safe to handle, but that is no reason to let your guard down as all of the devices are plugged into the ac mains. It also pays to remember that most integrated circuits of today are sensitive to static electricity, don’t handle them without grounding the static charge from your body.

INDEX TO POSTS UNDER CONSTRUCTION

How a computer works
What is a server?
Networking basics
Components of a Network
How it connects together
The infrastructure :: Network wiring and termination
Crimping and punching
Configuring the network
Cheap Hacks

Using apt-cacher to reduce your bandwidth consumption

Benjy — Tue, 29 Mar 2011 11:20:27 +0000

We had been managing on an Airtel 512kbps business plan for quite a while. When a new fourth office came up, it became apparent that our connection would no longer serve our purpose. For the fourth office, the Internet was critical and even after going in for a new 1Mbps leased line, our individual bandwidth did not seem to increase proportionately. It was time to crack the whip.

It was common knowledge that social networking, video sharing, news, photo and music sites were quite popular in office. These were soon blocked off either by file extension or by domain. However to be fair, there were other areas too where unnecessary bandwidth usage occurred. On analysis, two main areas turned up…

Peer to Peer file sharing
System updates

There were some systems where users who had administrative rights had quietly installed p2p software. There was nothing much I could do apart from reporting them as in an organisation of our kind it is impossible to strictly enforce an IT policy.

http://localhost:3142/ is where you can check whether apt-cacher is running...

Coming to system updates, this was one area which would make a huge difference to our Internet usage. Computers would needlessly download the same updates over and over again. In addition, many of the systems -especially laptops were set-up initially in our office, before being sent to the respective “branch offices”. This ended up in each fresh system downloading an incredible amount of patches, updates, fixes and service packs. For development too I needed the packages to be cached. Last month I installed Ubuntu server more than twenty six times on our new IBM x3650M3 server while trying out various experiments and each time needing to download all the updates in addition to the required packages and their dependencies. True I could have used a VM and restored the image, but I was messing with IBM’s serveRAID stuff and LVM as well.

We have the big three on our network and each of them are equally guilty of huge updates

Windows ( XP to 7)
MacOS 10x
Linux (Ubuntu desktops and servers)

I immediately dropped out the Mac and Windows from the list. The Mac thingy would need a Mac server and Windows would need at least a Windows 2003 server for running WSUS services not counting all the CAL’s I would need -one per client. I didn’t even bother to ask, as my organisation would think it trivial to cough out money for something (updates) that was happening anyway…

Apt-cacher generates a report every 24 hours which is available at http://localhost:3142/apt-cacher/report This shot was taken exactly after one day of usage.

I turned to Linux, which always has a cheap or free answer for everything. I imagined I’d have to setup a server and host repositories on it. I remembered what a nightmare it was when I tried it the last time. I had to have reps for two or three releases of Ubuntu with reps for both 32 and 64 bit versions and had to download all the packages and keep them updated. It was a veritable nightmare as our machines did not need half or more of the packages. It was an even greater loss of bandwidth for us.

Fortunately I came across apt-cacher. Caching is even better. It caches only what you need to download anyway. Installation and configuration is pretty straight forward. I did this on my desktop (actually an IBM tower server running Ubuntu Maverick)

sudo apt-get install apt-cacher

Next we need to edit the /etc/apt-cacher/apt-cacher.conf

There are lots of options here and they are documented well in the configuration file itself.
Here is a what I did.

First I set generate reports to 1

generate_reports=1

This generates a report every 24 hours which you can access from http://localhost:3142/apt-cacher/report [ Replace localhost with the IP address of the machine running apt-cacher ]

If you want to,you can import the existing packages in apt's cache on the system
it is running on into apt-cacher. This saves some unnecessary downloads.

sudo /usr/share/apt-cacher/apt-cacher-import.pl -s /var/cache/apt/archives

In addition you could pop in your Ubuntu distribution(s) and import all the files to save apt-cache from downloading them from the net. Files which are no longer referenced by the indexes get automatically cleaned up every 24 hours so that unnecessary packages are deleted and space on the disk is freed.

sudo /usr/share/apt-cacher/apt-cacher-import.pl -R -r /media/cdrom

Next I set apt-cacher to start automatically. Edit /etc/default/apt-cacher and set autostart to 1

AUTOSTART=1

Then, start the service manually for the first time, rather than restarting your system unnecessarily.

sudo service apt-cacher start

Entering the following url in your browser should show you that apt-cacher is running

http://localhost:3142/ [ Replace localhost with the IP address of the machine running apt-cacher ]

Finally you need to setup the clients to connect to the machine for updates. You can do this for the machine acting as the server as well.

Open a terminal window on each client and type sudo nautilus

Browse to /etc/apt/sources.list and right click and open it in your text editor

Press Ctrl H for the find and replace box.

In the find box, type http://

In the replace box, type http://192.168.1.50:3142/

Click on Replace all, Save and exit

You can now do an update and upgrade on the client. You can see the cached files in /var/cache/apt-cache/ on the server as the report will be generated only after 24 hours.

In case you see some updates not connecting to your server, these are probably listed under other software in the GUI based software sources editor. This can be accessed either through Synaptic or through Update manager. You need to select and edit each source manually. The entries there are usually the Dropbox, Google, Remastersys and any other third party repositories such as the Canonical’s Partner repositories.

Review: D-Link DNS-323 NAS Box

Benjy — Wed, 02 Feb 2011 15:44:42 +0000

NAS is finally getting cheap enough to fit the budget of big budget home owners and SOHO and SMB’s. With cameras, video recorders, movies and music all going digital, users now have huge amounts of data they want to protect optimally. With increased hard drive space being available at lower costs, redundant backup is slowly moving from server rooms to homes and small offices. Although NAS cannot handle the block level access like SQL databases need (For which a better option is DAS or SAN) it still makes sense as it is cheaper, reliable, easy to configure and less power hungry than using a server or Desktop computer as a file server for backups. It also has a smaller footprint and lower power consumption than a PC or server acting as a redundant file server. The Dlink DNS 323 has the footprint of an APC 600VA UPS and so does not take up much space on the desktop.

The Boxed DNS-323 and two 1TB hard disks (bought separately)

Last week I got an opportunity to source and setup a D-link DNS-323 NAS box. We sourced the box for around INR 9800/- and had the option to go in for 2 SATA disks of either 1TB or 2TB. The 2TB x 2 option would have cost us an extra INR 12000/- and the 1TB x 2 option about INR 6000/- extra. Both the prices were quite good taking into consideration all the advantages of the DNS-323. We bought the box for individual user data backups, so redundancy was important to us, so we configured the box in RAID1 (mirroring mode). The DNS-323 claims that when a faulty disk is replaced, it will automatically rebuild the array.

The DNS-323 supports the following disk (SATA only) configurations. (Assumption is that we are using 2 x 2TB disks)

JBOD (Linear) mode which will give us a disk of the total size of both disks (disk a + disk b) = 4TB Volume

RAID 0 Gives us a high performance 4TB volume

RAID 1 Gives us a 2TB redundant volume

JBOD and RAID 0 both offer no redundancy, so it makes sense to go in for RAID 0 as it offers better performance while giving you the same total volume size -unless your disks are of dissimilar size as the formula for calculating the array size for RAID 0 is 2 x (size of the smallest disk). The other advantage that JBOD has is that in case of a crash, data recovery would be relatively easier and cheaper.

The network interface on the DNS-323 is 10/100/1000 Gigabit ethernet. It also has a USB port which can be connected to a printer and allow it to act as a print server. It runs a flavor of Linux which supports both the ext2 and ext3 file systems. It helpfully advises you when configuring your disks that choosing ext2 will give you faster access times, while ext3 while slower, will store your data more reliably. It does not support ext4. It has a long list of protocols it supports, a few of which I am listing here.

Adding a disk drive to the NAS box is as easy as pushing it in

USB 2 Print Server, UPnP AV Media server for streaming to music, photos, and video to compatible media players, playstation 3 and Xbox 360, iTunes server, FTP Server etc..

Some of the software provided are, Dlink Easy Search, Drive clone pro etc.. This software enables manual backup and automatic backup as well. To be fair, I cannot comment on any of them as I’m a bit wary of bundled software. Many a time in the past, updates have not been provided and I don’t want to be shocked in the future when the next release of Windows comes by and the software proves truant or asks me to purchase a paid upgrade.

Close up of one of the drive bays.

Installing the hard disks was shockingly simple. The front of the box slides upwards and out like the battery cover on my Nokia 6275i. The hard disks are slid in and lock in place. They can be released by quick release levers behind the box. Put the cover back, connect it to a UPS and power it on and it starts to format the drives by itself. Of course once you configure the hard disks from the web interface, it formats the array again. In addition to the blue illuminated square that also functions as the power switch, there are 3 blue indicators on the bottom row. The center one shows network activity and the left and right one shows the state of the hard disks. The manual says that if a hard disk fails, the appropriate light will turn orange.

I did use the Dlink Easy search software as no matter what I did I could not locate the DNS-323 on the network. It was set to DHCP! Configuring the DNS-323 through its web interface was a breeze. I was able to create users and groups and assign different rights to either users and groups. I was able to turn on quotas for the users too. I setup a ftp server (supports sftp too) to receive the backups as SMB has too many protocol overheads. The reason I chose ftp over sftp was because the backups would take place over a private network and also because the freeware version of the backup software I was using would not support sftp.

Easy Web based administration

The DNS-323 supports a maximum of 10 simultaneous ftp connections. To be on the safe side, I staggered the backups and scheduled the backups to take place at half hour intervals on each client. I was happy with the speed at which the DNS-323 responded. The DNS-323 was located in my office, was switched to another office with a STP cable close to 100 meters in length, switched again and sent to a floor below, finally switched twice before it terminated at a 108MBPS access point. Not a very good way to setup a network, but our offices grew that way and some of them are temporary.

The best part is that I was able to setup the DNS-323 to email me the daily system and access log files and also a report detailing the amount of disk space used and free space remaining. The backups are going great! I’ll be shortly documenting the client side of the backup solution.

The NAS Box installed between a Netgear Access point and APC RS 600 UPS

The D-link DNS-323 is a great buy! Buying one for my home and connecting it to a wireless access point in client mode, I would have something better than Apple’s Time capsule!

-infinite control over what I could do with it, multi operating system compatibility, RAID, and of course no limitations enforced upon me by by Apple

Note: (20th April 2011)

We bought a new NAS box for our office for our accounts backup. The hard disks that we got this time were slim hard disks. This is when I noted that they were incredibly difficult to insert as they would not align with the SATA connectors on the rear as the bays were sized for thicker hard disks. Even after fitting, these hard disks keep wobbling around and might damaged the connectors if the box receives as sharp jolt. Dlink needs to work on bays that will fit both sizes of hard disks optimally.

Experiments with Meths

Benjy — Sun, 18 Jul 2010 08:12:02 +0000

Do a quick search of the forums and you find that the general opinion is that finding meths in India is hard. Well this may partly be true, because medical spirit or Isopropyl alcohol (they are not the same thing) is pretty expensive if you plan to use it for cooking. Ideally of course it would be wonderful to have your own reflux/moonshine still and burn pure ethanol in its virgin state, but that might probably also land you behind bars in this age of patriotic slavery. So we are stuck with burning methanol [ CH3OH ].

When mixed with water, it turns milky. this is the 1:1 which did not burn!

Methanol can be produced by the destructive distillation of wood (Pyrolysis), or from natural gas. It is also naturally produced during metabolism by many anaerobic bacteria.

Methanol is toxic. Wikipedia says that 10ml is sufficient to cause blindness by damage to the optic nerve and 30ml is fatal. On inquiring on Facebook, friends suggested that the paint & hardware shops were the right place to find it. I asked another friend and he said that some shady doctors in the slums bought it from there as it cost Rs. 50-55 per liter as compared to approximately the same cost for about 200ml of medical spirit. I bought 5 liters of the stuff after making sure that it was meths and not turpentine. You need to ask for spirit. If you ask for methylated spirits, meths or denatured alcohol, they wont have a clue what it is. A quick test is to pour some on your palm. It feels cool and evaporates completely and leaves your hand slightly moist -and leaves no deposits (very important).

At 1.5:1 The fuel mostly burns in the well. It can still cook, but it would be time consuming -especially in cold weather.

Since this was bought at a paint and hardware store, it was obviously not designed to be burned as fuel in a stove. (I’m not sure what it is used for -probably as a thinner for paint?) I’ve heard that the meths abroad are colored, just like our kerosene here is stained blue. This is pale yellow, like a light beer, which is probably its natural color. Since it wasn’t designed to be burned in a stove, I wasn’t surprised that it burned with a sooty bright yellow flame when the burner became hot. You can see examples of the various stages of burn in a previous post. At a certain point, the flame would flare up so much that I would drop the simmering ring onto the burner fearing that the burner would explode. After searching for pictures of a Trangia burner being used for cooking, I found many pictures of the burner with large yellow flames (unlike those lovely blue gas ring type of flames in most of the pictures) which has put my heart at ease, yet when I wanted to put out the flames it was difficult to extricate the hot simmer ring, slide the opening shut and drop it back onto the burner.

2.5:1 - The right mix -at least during the North Indian summers... The flame increases slightly and turns yellowish when the burner further heats up.

With all these things floating around in my mind, I remembered reading that meths are miscible with water and adding 10% of water to the meths causes it to burn with less soot. It also occurred to me that the diluted meths would be less volatile and burn gentler, so I decided to give it a try. I started off with a 1:1 meth:water ratio and went up to 3:1.

Some Observations…

The Meths turned milky on mixing it with water, like toddy or rice water
It is better to mix the meths and water outside the burner or whatever is added first gets wicked up the walls of the burner
Use distilled or soft water. If you use hard water, you can see a layer of salt floating in the fuel well making a cracking noise. This can be poured off, but anything that gets wicked into the walls will solidify there and become s a permanent part of your burner which is not so good!

Overkill at 3:1 This is just 30 seconds after lighting the burner

The results… [ Meths : Water ]

1 : 1 The meths-water mixture would not ignite even at an ambient temperature of about 40 deg C

1.5 : 1 The mixture ignited but burned mostly in the fuel well. there wasn’t enough fuel vaporizing to power the burner

2 : 1 Like above, there wasn’t enough fuel vaporizing to produce jetting, but intermittent jetting happened towards the end of the burn cycle as the Trangia burner heated up

2.5 : 1 Took some time for the jetting to begin, and started off with a nice blue flame, but turned into a gentle yellow flame which stayed the same even after the burner was too hot to touch. Ideal mixture at a room temperature of 40 deg C and strong enough not to go out even when I pointed my pedestal fan at it .(at its lowest speed)

3 : 1 Jets faster than the above mixture, but emits a tall yellow flame… although, it was not tested long enough, it is possible, that the flame would have got larger as the burner further heated up and vaporized more of the meths. There was no point in further testing as the previous 2.5 : 1 mixture proved ideal.

Tthe ideal 2.5:1 mixture needs to be tested at various ambient temperatures. I suppose it will take too long to jet at cooler temperatures and hence increase the cooking time. I’m traveling to Landour, Mussoorie on work tomorrow. Next weekend a few friends plan to camp somewhere in the hills. It should be at least 50% cooler than Delhi and probably raining. Good place to conduct additional tests. Will post the results on my return.

Caution: Check with the authorities before traveling whether you can carry meths with you in India. It is generally forbidden to carry any kind of inflammable goods in public transport whether by bus, train or air…. I still do, because I carry it in Trangias TUV certified safety bottle, but that won’t help me if I’m caught…

26th July 2010

Finally managed to test the 2.5:1 mixture at Landour in Musoorie. This was done close to the highest point of Musoorie (Lal Tibba) at an ambient temperature of 18 to 20 deg Celsius. It was the monsoon season and clouds were passing through the air. It did ignite here, but as a meth flame isn’t easily visible in the sunlight, I had to redo the test in my room at Landour Community Hospital. As expected, the burner took more than 15 minutes before finally jetting which is not acceptable, so a 3:1 mixture would probably have been more suitable. In the winters when it snows here, I’d probably need 4:1 or use undiluted meths. So there is no ideal mixture here. I will have to keep a chart of the temperatures and ratios.

22nd August 2010

Medical Spirit and white spirit

I was lucky to be able to get this picture. On the left is the medical/surgical spirit that is available to hospitals and medical practitioners for their use (For about Rs. 70/- or so per liter [at wholesale rates] ). I think this is the spirit available overseas which is sold as meths. I remember other bloggers in the UK writing about blue meths. Next to the can of blue meths, is a bottle of white spirit. This is probably 100% Ethanol and as clear as water. It can be allegedly drunk after being filtered through charcoal and is also what booze companies use as an additive to their products. I also heard a story of someone caught in Bangalore (now Bengaluru) selling “Brandy” which was actually black tea mixed with this stuff. This should be very hard (and illegal) to obtain. You would need the right connections to obtain this stuff. It would be lovely for burning though! This bottle came for a government funded medical project. I’m sure there are a lot of unscrupulous fellows siphoning this off to local bootleggers as there would be definitely “good” money in this. Sometime in the future, I must try my hand at distilling my inferior yellow meths and see whether I can get white spirit good enough to burn… I suspect, using what I currently use would block my Trangia burner with sediments in the long run!

Simplysimple.info » DIY

Components of a Network

Related Posts

How it connects together

Related Posts

Networking basics

Related Posts

What is a server?

Related Posts

How a computer works

Related Posts

Networking Simplified

Related Posts

Using apt-cacher to reduce your bandwidth consumption

Related Posts

Review: D-Link DNS-323 NAS Box

Related Posts

Experiments with Meths

Some Observations…

Overkill at 3:1 This is just 30 seconds after lighting the burner

The results… [ Meths : Water ]

26th July 2010

22nd August 2010

Related Posts

Simplysimple.info » DIY

Components of a Network

Related Posts

How it connects together

Related Posts

Networking basics

Related Posts

What is a server?

Related Posts

How a computer works

Related Posts

Networking Simplified

Related Posts

Using apt-cacher to reduce your bandwidth consumption

Related Posts

Review: D-Link DNS-323 NAS Box

Related Posts

Experiments with Meths

Some Observations…

Overkill at 3:1 This is just 30 seconds after lighting the burner The results… [ Meths : Water ]

26th July 2010

22nd August 2010

Related Posts

Overkill at 3:1 This is just 30 seconds after lighting the burner

The results… [ Meths : Water ]