Bryan Valentine

Kevin Harrington

LIS4366 Web Development

16 April 2006

Project 1-1

Currently, there are 386 recognized distributions of the Linux system.

Conectiva Linux is a mainstream, Intel compatible, multi-language system. Mainstream mostly means that this distribution is meant for the general public. This distribution was created in Brazil is contains all the usually advantages of mainstream distributions. These include ease of installation and setup. There edition for workstations come standards with extra software packages that include StarOffice, Netscape (native languages), game, image manipulation applications, and a few other utilities. The other version titled the Server edition also comes with software. Some of these include network administrations and e-commerce tools. Conectiva uses a test certificate. In a production server a signed CA certificate must be used.

Beehive Linux is not a mainstream distribution but instead, it was designed for the minimalist user. This distribution of Linux is Intel compatible, but is only available in one language, which is English. This version is labeled as a scaled down version that promised to take out all the clutter of other distributions. This software should run tighter and cleaner, given its precision workings. This distribution is recommended for experienced Linux users.

Hard Hat Linux is an embedded, non-Intel compatible distribution of Linux. Hard Hat and CDK are built from 100% pure Linux components and distributed as an open source tool kit. MontaVista (creator) ships both source and easy-to-use pre-built binary packages on the same media. Hard Hat Linux includes no hidden proprietary components, and is always completely free. Hard Hat Linux Cross Development Kit (CDK), targets the broadest array of CPU architectures and board and system-level platforms of any embedded Linux development solution available today. Hard Hat Linux to provide robust Linux kernel ports, device drivers, middleware, and development tools to streamline your development effort.

The use of Conectiva would be the best choice for me in using Linux for a web server. Beehive is for an experienced user which I’m not. Hard Hat is not Intel compatible. I will be using a Intel based system, therefore Hard Hat is not for me. Conectiva is a simple to use mainstream distribution that many people use. I feel that I would get lots of help if needed. With the extra software included in the package like network administrations tools, and the use of different languages, Conectiva is the best possible choice.

Project 1-2

Kentrox Q2200 T1 CSU/DSU Router Combining the features of an IP router, QoS appliance, IPSec VPN appliance, firewall, T1 DSU/CSU, and Ethernet switch, the Q2200 QoS Access Router saves you money right out of the box. Setup and maintenance become far less time-consuming and expensive with an integrated device. And the Q2200 router's intuitive graphical user interface (GUI) helps remote offices get up and running quickly

IPLink 2603 T1 Router Combining ease-of-use with a full suite of LAN/WAN routing features, the IPLink routers provide selectable bridging or routing functionality along with advanced IP features such as NAT/NAPT, Firewall, and DHCP. A complete set of configurable FR/PPP/IP WAN protocols allow a wide range of choices when connecting branches via common WAN services. The IPLink Routers boast easy installation offering Console/VT-100, Telnet, and HTTP/SNMP management options. All IPLink routers come with an auto-sensing full-duplex 10/100Base-T Ethernet port, cross-over switch, and internal power supply. The Model 2603 is equipped with an integrated T1/E1 CSU/DSU for connection to full and fractional T1/E1 services.

Mega Path Full T1 Service MegaPath T1 offers reliability, without the high cost traditionally charged by phone companies. Businesses with T1 service can save hundreds of dollars per month by switching to MegaPath. Our T1 Internet access service offers 1.5 Mbps of always-available Internet bandwidth at a low, fixed monthly price. Plus our T1 service is backed by an industry leading SLA and guaranteed superior customer service!

Project 1-3

Currently in my area, which is, 2700 W Pensacola St Apt 126 Tallahassee, FL 32314, only a few options are available. The first option is from a company called Cybertronic, and they offer SDSL. An SDSL is a single digital subscriber line, which means it requires only one line to operate. Other earlier DSL types required more than one lines to operate. The main advantage of a SDSL is that the upload and download speeds are the same, at around 128kbps. This service will cost me around 119$ per month.

The same company offers ADSL at the same location. ADSL is known as an asymmetric digital subscriber line. This is for the typical home user that is interested in fast downloads speeds. The upload speed is less than the download speed by a good bit. The speed offered by Cybertronic at my location is 1500kbps. This is at 69.99$ per month, after the purchase of equipment.

A local company partnered with Cybertronic provides IDSL to my area. ISDN DSL is very limited. It works at one speed only 128kbps for uploads and downloads. This is provided for homes that are out of range. This service will cost around $119 per month.

The MegaPath Company provides to my location as well. They provide all the same services as Cybertronic, but without IDSL. Being IDSL isn’t a very desirable service; I don’t feel I will be missing anything. Their prices range from 150$ to 250$. After reading a few reviews I would decide not to use MegaPath services, based on poor customer services, and complicated setup.

The last company Direct Internet Access is partnered with a well known company called Bellsouth. They offer all of the same services IDSL, ADSL, and SDSL. The reviews were all very good. The pricing for Bellsouth is normal around $99 a month of ADSL. All the applications from Bellsouth offer free hook-ups. This service come with 10 mail boxes, and is around 1.5 mbps download.

Project 1-4

Microsoft Exchange Server 2003 Features include support for mobility and message hygiene. Improvements have been made to address performance and reliability with public folders and with database size restrictions.
Exchange maintains an open connection between the mobile device and the server. Remote wipe is another new feature, and it enables administrators to delete sensitive data from a lost or stolen mobile device. Other features include global address list (GAL) lookup, policy provisioning whereby administrators can make supported policies more secure, support for certificate-based authentication, use of S/MIME to sign and encrypt mail, and server-based synchronization of Tasks.
Anti-spam improvements are included in the release of the integrated version 2 of Microsoft Exchange Intelligent Message Filter, and Sender ID, which is an industry-standard framework. Version 2 of Intelligent Message Filter contains significant improvements in the anti-spam area for SP2.
For Exchange Server 2003 Standard Edition, the hard-coded licensing database size limit has been increased from 16 GB to 75 GB. The administrator can set a protective database size limit (prevent unintentional database size growth). The default value will be 18 GB in SP2 for Standard Edition and the default can be changed.

Microsoft Virtual Server 2005 This Server is an ideal, production-quality tool for consolidating multiple workloads onto a physical server, allowing organizations to make more efficient use of their hardware resources. Built upon Microsoft Virtual Server 2005 SP1 and bolstered with new functionality, Virtual Server 2005 R2 enables IT organizations to enhance their administrative productivity and rapidly deploy new servers to address changing business needs through automated deployment and configuration of connected virtual machines that are easily administered with standard server management tools. Virtual Server 2005 R2 is an extensively tested and well supported virtualization solution that is supported both by the broader ISV community and by Microsoft in conjunction with its server operating systems and applications.

Project 1-5

This project was completed at my home computer in a LAN at an apartment complex. Upon typing in tracert www.yahoo.com I got this message, “Tracing route to www.yahoo-ht2.akadns.net at IP address: 69.147.114.210.” From this point the prompt told me there would be a maximum of 30 hops. Only one hop was noticed. This hop was from router 192.168.1.1. Other data is present, but I don’t know what it means. All other possible hops are timed out.

The next test I did was to trace the website, www.linux.org. This trace provided mainly the same results as the other trace. The information I got was, “Tracing route to www.linux.org at IP address: 198.182.196.56. This too showed that there would be a maximum of 30 hops. The first hop took place, and it was from the same router as before in the other trace. This IP address was 192.168.1.1. There were many other hop attempts but they were all timed out. Finally at the end I had one last hope and that information said, “118ms 128ms 1126ms www.linux.org 198.182.196.56.”

Once I was able to do this trace at my school computer, I found that I had similar results. For www.yahoo.com I still was accessing the same server and IP address. However, the router was different. The router address on the school computer that I was going through was vlan.971.c1.dif.fsu.edu. This makes sense to, because the building that I am currently in gets its Internet connectivity through the Diffenbaugh (DIF) building on campus. There was only one hop. It tried a maximum of 30 hops but all the others timed out.

The same results were recorded when I did a trace of www.linux.org. The only change was that instead of looking at the yahoo.com server, I was looking at the Linux server. The routing was just the same.

Assignment #2

Project 2-1

Most Economical / The Best
Processor / AMD Athlon 64 / Pentium 4
Disk / WD Caviar SE16 / WD Raptor
RAM / Crucial 512MB DIMM / Crucial 1GB DIMM
NIC / Intel Pro 100 S / Intel Pro 1000 PT

When it comes down to processors, AMD Athlon chips rank in the most economical column. When I went to buy a computer, I was given mixed reviews, but the common theme was that for less money, I could get an AMD chip that would run possibly just as good as a Pentium 4. Intel has been making chips for years, and the Pentium 4 is their all around best chip. Looking at hard drives, the two important specifications are rotation speed, and size. For the economical column, I chose a Western Digital 7200RPM 250GB hard drive. In the best category, I chose the WD Raptor, which includes 150GB of space, but runs at 10,000RPMs. As far as economical, RAM should be around the 512MB range. This is about an average baseline level of memory. On the best side, moving up to a 1GB memory stick or multiple 1GB memory sticks would be worthwhile. Finally, the Network Interface Card (NIC). Based on economics, the Intel Pro 100 S is a good choice with 100mbps. This is a good affordable starting point for a NIC. To upgrade to the best, would be using an Intel Pro 1000 PT, which is 1000mbps.

Project 2-2

Because the environment is used constantly, speed of backups is important. Actually, in this setting speed is really the only factor. I would select between Differential backups and Incremental backups. Both of these procedures are most time saving then say full backups. I like the idea of Differential backups because instead of making a full backup, the backup only applies to new data since the last full back up. This way, there is less time wasted. Besides, it seems pointless to keep backing up information that has not changed a bit. Differential backups are logical. Full backup are still required, but as a company, there services most likely have certain periods of the day in which traffic has declined enough that a long backup process can take place. The main disadvantage to this procedure is that even though a full back up has been made, and all new backups are of changes, each new change has to be backed up. Every time there is new data, the new data, and the other data that was not on the full backup, gets backed up. This procedure is a much better idea, in that I have seen a server need to take two to three straight days to back up. This was a lot of data to backup, but what company is going to have hardly any data to backup?

The other procedure I mentioned earlier called Incremental, is similar to differential backups. Once a full backup has been taken all other new information gets back up (incremental backups). From that last incremental backup new information needs to be backed up. What makes incremental backups so nice is that each new backup is based on the last incremental backup. This way, you never have to backup data more than once until your next full backup. Like most things in life there is a disadvantage to incremental backups. While using incremental backups there is no copied data. If data needs to be restored then the user has to look back in any of the following: full backups, incremental backups (which could entail many backups). Deciding between the two procedures listed here is difficult. I feel that both are helpful, but each has a nice feature that the other doesn’t have. More than likely for a huge company because time is so important, differential backups would probably be used.

Project 2-3

Fault tolerance means the ability to of a system to keep running even when a component fails. Companies such as Amazon.com would lose substantial amount of money in the event that there e-commerce server’s crash. Many methods can be applied to have a high fault tolerance.

The first of these methods is called RAID 1-5, 10. RAID means achieving disk redundancy. A RAID system can prevent data loss when a single drive malfunctions. RAID 1 is when data is duplicated across two separate drives, but this takes more time to write data, being it needs to be written twice. This technique can also be called disk duplexing. RAID 2 uses a special error-correcting code for drives. This allows the drive to fix itself from minor glitches and what not. RAID 3 which is not a common methodology stripes data across the disk drives at the byte level. All information needed to rebuild data is saved on one disk. RAID 4 is very similar to RAID 3. The only difference is speed. RAID 5 is the most common disk redundancy used today. RAID 5 distributes data across the drives and stores the redundancy. This is important because if a disk does fail that it can be replaced with no loss of data. This RAID method is actually affordable too! RAID 10 is built on the same concept as RAID 5 but instead of using multiple single drives, it uses and array of mirrored RAID 1 drives. This is good because it eliminates the need to replace a drive like with RAID 5. The only drawback to RAID 10 is price!