LeeWare Self Managed Linux Dedicated Server Hosting

About LeeWare

Affordable Dedicated Servers for the Experienced Administrator

Are LeeWare Servers Right For Me?
This service isn't for everyone because LeeWare strives to maintain a "clean network", the following activities are not allowed: (1) No IRC (2) No Proxy Servers (3) No File Sharing or P2P (4) You must request a waiver before any email servers will work (no spam) (5) Limited support: All servers are deployed in unstaffed remote computing facilities. This has a variety of implications therefore, you should really be comfortable managing a machine remotely. If you have issues, you need to send an email to admin and it will be addressed during a maintenance window (typically, twice daily). Facilities issues i.e. things outside of your control i.e. power issues, network, HVAC etc. result in auto notifications to the admin ETAs are 1-2 hours however, typical times are 30 minutes or less. (6) default Operating System: CentOS 5.x x64 / Ubuntu Server x64 (available upon request 40 minute install per system). I can deploy Windows Servers but the customer must supply a Valid License Key to activate + ISO image. + Pay a $25 Per machine setup fee and $25 rhands fee per reload. Windows boxes have the longest setup and patch times i.e. 2 hours. (7) Customers must supply and make arrangements for their own Control Panels. (8) No hardware modifications - I can provide customers with additional servers if they need to increase capacity at a discount. (10) Urgent issues which require a site visit be conducted outside of normal hours will require a $25 rhands fee prior to dispatch. Abuse issues will result in corrective action including service termination. (11) This service is primarily intended for the hosting and horizontal scaling of CPU/MEM/DISK intensive interactive web applications. Not passive bulk data applications like streaming media or content delivery. Therefore, if you need to host high bandwidth applications you should try http://www.100tb.com or http://www.fdcservers.net. (12) This service is intended to be used by competent systems administrators who know what they are doing and more importantly why. (13) If you are planning to resell this service, then you need to make sure your basic offerings do not promise your customers more than you are paying for and are offer directly in terms of service. Furthermore, you need to make sure the customers you put on the system meet the desired profile type suitable for this service. Failure to adhere to this will lead to conflicts and possibly termination primarily because this service isn't structured to support the demands of general purpose resellers. For example, my time is very expensive, no ones service fees covers my travel expenses and time. Therefore, I cannot spend every maintenance period (intended to physically check environment + make service adjustments) running from site-to-site to reinstall a bunch of machines and reboot others because the sub-customers are screwing them up or because as a reseller you can't make up your mind about what OS you want to offer your customers. (14) Your expectations should be aligned with the prices you are paying. This service is intended for people who need "good enough" reliability and service. Which means that this service is primarily targeted towards a niche market and not WHT types.

Are LeeWare Servers Right For Me?

This service isn't for everyone because LeeWare strives to maintain a "clean network", the following activities are not allowed: (1) No IRC (2) No Proxy Servers (3) No File Sharing or P2P (4) You must request a waiver before any email servers will work (no spam) (5) Limited support: All servers are deployed in unstaffed remote computing facilities. This has a variety of implications therefore, you should really be comfortable managing a machine remotely. If you have issues, you need to send an email to admin and it will be addressed during a maintenance window (typically, twice daily). Facilities issues i.e. things outside of your control i.e. power issues, network, HVAC etc. result in auto notifications to the admin ETAs are 1-2 hours however, typical times are 30 minutes or less. (6) default Operating System: CentOS 5.x x64 / Ubuntu Server x64 (available upon request 40 minute install per system). I can deploy Windows Servers but the customer must supply a Valid License Key to activate + ISO image. + Pay a $25 Per machine setup fee and $25 rhands fee per reload. Windows boxes have the longest setup and patch times i.e. 2 hours. (7) Customers must supply and make arrangements for their own Control Panels. (8) No hardware modifications - I can provide customers with additional servers if they need to increase capacity at a discount. (10) Urgent issues which require a site visit be conducted outside of normal hours will require a $25 rhands fee prior to dispatch. Abuse issues will result in corrective action including service termination. (11) This service is primarily intended for the hosting and horizontal scaling of CPU/MEM/DISK intensive interactive web applications. Not passive bulk data applications like streaming media or content delivery. Therefore, if you need to host high bandwidth applications you should try http://www.100tb.com or http://www.fdcservers.net. (12) This service is intended to be used by competent systems administrators who know what they are doing and more importantly why. (13) If you are planning to resell this service, then you need to make sure your basic offerings do not promise your customers more than you are paying for and are offer directly in terms of service. Furthermore, you need to make sure the customers you put on the system meet the desired profile type suitable for this service. Failure to adhere to this will lead to conflicts and possibly termination primarily because this service isn't structured to support the demands of general purpose resellers. For example, my time is very expensive, no ones service fees covers my travel expenses and time. Therefore, I cannot spend every maintenance period (intended to physically check environment + make service adjustments) running from site-to-site to reinstall a bunch of machines and reboot others because the sub-customers are screwing them up or because as a reseller you can't make up your mind about what OS you want to offer your customers. (14) Your expectations should be aligned with the prices you are paying. This service is intended for people who need "good enough" reliability and service. Which means that this service is primarily targeted towards a niche market and not WHT types.

Understanding Pooled Hardware Selection Process
There are one-thousand servers in the hardware pool. All packages are put together by selecting idle hardware in this pool . It is very important that you understand how this pooled hardware selection process works. This hardware is generally similar in terms of performance and CPU clock rates (2.X Ghz+) but in the interest of full disclosure, the machines might not be identical. (1) The current production set of hardware is all AMD-64 Based. (2) This includes Single Core AMD-64 Athlons, AMD-64 Semprons LE-1250 (G2): Multi-core AMD-64 Athlon X2, AMD-64 Phenom X3 and finally AMD-64 Phenom X4. (3) There is a split between the AMD Athlon based hardware and the AMD Sempron based systems and that is, the AMD-64 Athlon are socket 939 and the motherboards support a max memory configuration of 4GB resulting in a memory hole which makes only 3.2GB usable. The rest of the systems are all based on AM2 technology and also support a max of 4GB of memory However, these don't have the memory hole issue. (4) Hard drives will be one of the following: (based on availability) 200GB/ 250GB Larger or slightly smaller hard drives will be used if availability is limited (160GB) (5) The pool selection works as follows: All orders are placed with the minimum hardware specifications i.e. AMD-64 Single Core Systems. The selection algorithm: Fill order with AMD-64 Single Core Systems, If no single core systems are available or there are not enough to complete the order fill with X2s if there are no X2 fill with X3. X4s are currently locked out of the pool selection process. Therefore, it is not possible to dictate through the normal ordering process which systems you will receive. There are plans to fix this in the future but this will be reflected in the costs of the packages. The multi-core packages will have a much higher cost because they have much higher power and cooling costs vs the single core systems. Therefore, for the best value you can just roll the dice to see what you get you never know. Or if you require more certainty you can always check current availability by sending an email to sale or admin requesting a current availability report by indicating the number of systems you plan to order and I can tell you what your order will look like if you submitted a requests for systems.

Understanding Pooled Hardware Selection Process

There are one-thousand servers in the hardware pool. All packages are put together by selecting idle hardware in this pool . It is very important that you understand how this pooled hardware selection process works. This hardware is generally similar in terms of performance and CPU clock rates (2.X Ghz+) but in the interest of full disclosure, the machines might not be identical. (1) The current production set of hardware is all AMD-64 Based. (2) This includes Single Core AMD-64 Athlons, AMD-64 Semprons LE-1250 (G2): Multi-core AMD-64 Athlon X2, AMD-64 Phenom X3 and finally AMD-64 Phenom X4. (3) There is a split between the AMD Athlon based hardware and the AMD Sempron based systems and that is, the AMD-64 Athlon are socket 939 and the motherboards support a max memory configuration of 4GB resulting in a memory hole which makes only 3.2GB usable. The rest of the systems are all based on AM2 technology and also support a max of 4GB of memory However, these don't have the memory hole issue. (4) Hard drives will be one of the following: (based on availability) 200GB/ 250GB Larger or slightly smaller hard drives will be used if availability is limited (160GB) (5) The pool selection works as follows: All orders are placed with the minimum hardware specifications i.e. AMD-64 Single Core Systems. The selection algorithm: Fill order with AMD-64 Single Core Systems, If no single core systems are available or there are not enough to complete the order fill with X2s if there are no X2 fill with X3. X4s are currently locked out of the pool selection process. Therefore, it is not possible to dictate through the normal ordering process which systems you will receive. There are plans to fix this in the future but this will be reflected in the costs of the packages. The multi-core packages will have a much higher cost because they have much higher power and cooling costs vs the single core systems. Therefore, for the best value you can just roll the dice to see what you get you never know. Or if you require more certainty you can always check current availability by sending an email to sale or admin requesting a current availability report by indicating the number of systems you plan to order and I can tell you what your order will look like if you submitted a requests for systems.

Understanding Average Rate Bandwidth Allocations
(1) All bandwidth allocations are allocated to servers as a group NOT individually, Therefore, if you order a LWC10x5 [1500GB] you are ordering a server package of 10 dedicated servers with a pooled allocation of 5Mbit/s to be shared by all the servers in the group. Alternatively, you can order packages with pooled bandwidth allocations for 10Mbit/s, 20Mbit/s, 40Mbit/s or higher but that will dramatically increase the cost of your packages and I am pretty sure you won't like the prices. (2) This 5Mbit/s is not a hard limit but an average amount of bandwidth your servers should consume over the course of a single month. If you are a WHT type or someone with very limited administrative experience this will seam like a laughable bandwidth allocation. However, if you have any real experiencing hosting web-based applications you will realize that (most) 98% of web applications use between 500Kbit/s ~2Mbit/s and if the page size is 15KB (including images ) (120.000bits) after accounting for 15% over head with 5Mbit/s you can serve that page 35 times per second. Which means that if this was a constant stream of traffic and we know that most websites have peaks and valleys 1500GB/month would allow you to serve that page 104,857,600 times! Therefore, if you load-balanced this over 10 servers you would get 3.5/hits/users per second and about 10,000+ hits per server/month. (3) Because the 5Mbit/s is not a hard cap (you can burst) using the same aforementioned example your LWC10x5 package could handle serving that same 15KB page 708 times per second for a monthly total bandwidth utilization of 2.75TB or 196,852,668 times!. I can say with confidence that most websites / applications come no where near these numbers. Which leaves all of your high bandwidth users to the usual suspects. File Sharers, Streaming Media Sites, Proxy site operators which pretty much allow file sharing and streaming or file download sites via a proxy. (4) There are occasional people that sign up to this service either directly (rare) or through a reseller (often) and then attempt to run expressly prohibited network applications. In the past, a lot of this activity would lead to network disruptions because servers were completely uncapped and without any kind of traffic controls. It was possible for a single customer with an LWC20x5 to generate 2Gbit/s of traffic (that they have no intentions of paying for) and dumping that much volume on the Metro-Net with the normal traffic which might be 300Mbit/s. Starts to cause high-latency and dropped packets as the Gig E links become overwhelmed with junk traffic. Therefore, I have implemented global traffic control policies which will allow normal activity to work without issue, but expressly prohibited traffic will under perform. (5) A network accounting system is in place to detect and report expressly prohibited usage of the service. Customers who are flagged for such activities will be immediately dropped from the service.

Understanding Average Rate Bandwidth Allocations

(1) All bandwidth allocations are allocated to servers as a group NOT individually, Therefore, if you order a LWC10x5 [1500GB] you are ordering a server package of 10 dedicated servers with a pooled allocation of 5Mbit/s to be shared by all the servers in the group. Alternatively, you can order packages with pooled bandwidth allocations for 10Mbit/s, 20Mbit/s, 40Mbit/s or higher but that will dramatically increase the cost of your packages and I am pretty sure you won't like the prices. (2) This 5Mbit/s is not a hard limit but an average amount of bandwidth your servers should consume over the course of a single month. If you are a WHT type or someone with very limited administrative experience this will seam like a laughable bandwidth allocation. However, if you have any real experiencing hosting web-based applications you will realize that (most) 98% of web applications use between 500Kbit/s ~2Mbit/s and if the page size is 15KB (including images ) (120.000bits) after accounting for 15% over head with 5Mbit/s you can serve that page 35 times per second. Which means that if this was a constant stream of traffic and we know that most websites have peaks and valleys 1500GB/month would allow you to serve that page 104,857,600 times! Therefore, if you load-balanced this over 10 servers you would get 3.5/hits/users per second and about 10,000+ hits per server/month. (3) Because the 5Mbit/s is not a hard cap (you can burst) using the same aforementioned example your LWC10x5 package could handle serving that same 15KB page 708 times per second for a monthly total bandwidth utilization of 2.75TB or 196,852,668 times!. I can say with confidence that most websites / applications come no where near these numbers. Which leaves all of your high bandwidth users to the usual suspects. File Sharers, Streaming Media Sites, Proxy site operators which pretty much allow file sharing and streaming or file download sites via a proxy. (4) There are occasional people that sign up to this service either directly (rare) or through a reseller (often) and then attempt to run expressly prohibited network applications. In the past, a lot of this activity would lead to network disruptions because servers were completely uncapped and without any kind of traffic controls. It was possible for a single customer with an LWC20x5 to generate 2Gbit/s of traffic (that they have no intentions of paying for) and dumping that much volume on the Metro-Net with the normal traffic which might be 300Mbit/s. Starts to cause high-latency and dropped packets as the Gig E links become overwhelmed with junk traffic. Therefore, I have implemented global traffic control policies which will allow normal activity to work without issue, but expressly prohibited traffic will under perform. (5) A network accounting system is in place to detect and report expressly prohibited usage of the service. Customers who are flagged for such activities will be immediately dropped from the service.

cat /proc/cpuinfo wrong clock rate!
All CPUs in production are running at 2200Mhz (2.XGhz) However, on systems that support cpu-auto-scaling (a power saving feature) the CPU will run at a lower rate i.e. 1000Mhz (1.0Ghz) when the system is lightly loaded, a cat /proc/cpuinfo will reflect this. The clock rate of the CPU will automatically increase as system demand goes up. Therefore, if you run a processor intensive job and then issue a cat /proc/cpuinfo you will see the correct clock rate. If you want to verify the clock rate of the CPU on a system that supports cpu-auto-scaling then just issue the following command: #dmesg \| grep MHz

cat /proc/cpuinfo wrong clock rate!

All CPUs in production are running at 2200Mhz (2.XGhz) However, on systems that support cpu-auto-scaling (a power saving feature) the CPU will run at a lower rate i.e. 1000Mhz (1.0Ghz) when the system is lightly loaded, a cat /proc/cpuinfo will reflect this. The clock rate of the CPU will automatically increase as system demand goes up. Therefore, if you run a processor intensive job and then issue a cat /proc/cpuinfo you will see the correct clock rate. If you want to verify the clock rate of the CPU on a system that supports cpu-auto-scaling then just issue the following command: #dmesg | grep MHz

Accessing Your Server For the First Time
There are a couple of ways for you to access your server after signing up. You will receive a welcome e-mail containing your user name/passwords and server assignments. If you use Linux or Unix you can use Secure Shell (ssh) this service is similar to telnet but it encrypts the data between your computer and the server. To connect to the server from the command line, issue the command: ssh username@ServerNameOrIPAddress The server will prompt you for your password and you will have access to your console after supplying it. To disconnect, just type log out. If you are accessing your servers from a windows-based machine you can use PuTTY a free SSH client to access your server(s).

Accessing Your Server For the First Time

There are a couple of ways for you to access your server after signing up. You will receive a welcome e-mail containing your user name/passwords and server assignments. If you use Linux or Unix you can use Secure Shell (ssh) this service is similar to telnet but it encrypts the data between your computer and the server. To connect to the server from the command line, issue the command: ssh username@ServerNameOrIPAddress The server will prompt you for your password and you will have access to your console after supplying it. To disconnect, just type log out. If you are accessing your servers from a windows-based machine you can use PuTTY a free SSH client to access your server(s).

How to make support request
First, all support requests should be sent to admin@leeware.com Please include your server name in the subject of your email message along with your request. This information can be found in your welcome email e.g. cipher.leeware.com [reboot\| reinstall \| help \| question \| etc] Doing this will allow me to address issues more quickly.

Default Server Configuration
By default servers are configured with the following options - (1) Base Operating System with SSH enabled. (2) All servers were patched at install time. (3) You are responsible for subsequent updates. (4) To reduce the turn-around time and serve the needs of the vast majority of customers, the idle server Capacity is set to the latest version of CentOS (x86_64). If you want Ubuntu x64 installed on all of your boxes or some sub-set of them, you should indicate that prior to placing an order either by using the web-form or by sending an email to sales@leeware.com to indicate this. You should be aware that deviating from CentOS on your initial order can increase your setup times because depending which sites your servers are physically located plus the size of your order, you are looking at an install time of roughly 40 minutes per system where much of this will be squeezed in between other tasks. Therefore, it could take 24-48 hours to provision 10 server packages and 48-72 hours to provision 20 server or larger packages.

Default Server Configuration

By default servers are configured with the following options - (1) Base Operating System with SSH enabled. (2) All servers were patched at install time. (3) You are responsible for subsequent updates. (4) To reduce the turn-around time and serve the needs of the vast majority of customers, the idle server Capacity is set to the latest version of CentOS (x86_64). If you want Ubuntu x64 installed on all of your boxes or some sub-set of them, you should indicate that prior to placing an order either by using the web-form or by sending an email to sales@leeware.com to indicate this. You should be aware that deviating from CentOS on your initial order can increase your setup times because depending which sites your servers are physically located plus the size of your order, you are looking at an install time of roughly 40 minutes per system where much of this will be squeezed in between other tasks. Therefore, it could take 24-48 hours to provision 10 server packages and 48-72 hours to provision 20 server or larger packages.

Administering Your Server
(1) This is an Unmanaged Service therefore, you are responsible for installation, configuration and troubleshooting issues related to your server. (2) I will handle pretty much anything that you can't do for yourself like: Restarting a crashed server, Turning off a firewall [you should always include instructions on how to disable it.], network issues that might keep you from accessing your server if those issues originate from my end of the connection. (3) Regarding system backups: It is your responsibility to make and maintain them. (4) If a service conflict arises i.e. you cancel your service or default on a payment and lose access to your systems don't expect to get access without payment so that you can "download your files" or "make backups." Because when a cancellation is received, machines are disconnected and queued to be returned to the free systems pool and "downloading files" and "making backups" is something you should have done prior to cancellation.

Administering Your Server

(1) This is an Unmanaged Service therefore, you are responsible for installation, configuration and troubleshooting issues related to your server. (2) I will handle pretty much anything that you can't do for yourself like: Restarting a crashed server, Turning off a firewall [you should always include instructions on how to disable it.], network issues that might keep you from accessing your server if those issues originate from my end of the connection. (3) Regarding system backups: It is your responsibility to make and maintain them. (4) If a service conflict arises i.e. you cancel your service or default on a payment and lose access to your systems don't expect to get access without payment so that you can "download your files" or "make backups." Because when a cancellation is received, machines are disconnected and queued to be returned to the free systems pool and "downloading files" and "making backups" is something you should have done prior to cancellation.

Value for the money
The major value proposition for this service is simply there is a market for low-cost, affordable or budget based services. The challenge is being able to offer that service effectively which means part of the battle is managing the customer expectation. Let me talk a little bit about what makes LeeWare different from most (but not all) budget dedicated server providers. (1) This service is backed by a Professional Network Consultant and Systems Administrator. In the 20+ years I have been doing this work, I have planned and implemented several Corporate I.T. systems in a variety of (shocking) environments all across the US and what I learned from my Corporate customers is that they will pay good money for people and technology that helps them to meet their business goals and needs. What they don't like is wasting money on stuff that doesn't matter. Furthermore, they have a pretty firm grasp on their actual need vs. what they (think) they need based on some fuzzy notion or marketing hype. This service was born to provide best effort services to savvy systems administrators. (2) Service transparency: Due to the price point of my products I attract a fair amount of non-savvy customer types (WHT) that confuse my services with those being operated by kiddy-hosts and amateurs so I end up referring those customers to other budget providers who are setup to provide services for those client types. Because the prices are low doesn't mean that the service is inferior. It just means that you are trading a lot of non-essential frills and reliability for a lower price point. I can think of many multi-carrier, redundant this and that data centers for which customers pay a premium for only to end up suffering the same or if not more outages with those providers. For a catalog of this I will provide two references: [1] http://www.datacenterknowledge.com/archives/category/downtime/ and [2] http://www.webhostingtalk.com/forumdisplay.php?f=59 Therefore, for the price you are paying or the money you are saving , you should make appropriate investments to create your own redundency or tolerate the failures when they occur. (3) Due diligence: Because the main focus of my business is providing a platform for people to host applications and services. I have a lot of my money invested into this venture therefore, it is important to me to make that platform as reliable as possible within reason relative to the revenues generated by the service. Because increased system reliability equals increased customer confidence. Therefore, I am very proactive about keeping disruptions to a minimum. In fact, I think that you will find that this service has better than expected reliability for the prices.

Value for the money

The major value proposition for this service is simply there is a market for low-cost, affordable or budget based services. The challenge is being able to offer that service effectively which means part of the battle is managing the customer expectation. Let me talk a little bit about what makes LeeWare different from most (but not all) budget dedicated server providers. (1) This service is backed by a Professional Network Consultant and Systems Administrator. In the 20+ years I have been doing this work, I have planned and implemented several Corporate I.T. systems in a variety of (shocking) environments all across the US and what I learned from my Corporate customers is that they will pay good money for people and technology that helps them to meet their business goals and needs. What they don't like is wasting money on stuff that doesn't matter. Furthermore, they have a pretty firm grasp on their actual need vs. what they (think) they need based on some fuzzy notion or marketing hype. This service was born to provide best effort services to savvy systems administrators. (2) Service transparency: Due to the price point of my products I attract a fair amount of non-savvy customer types (WHT) that confuse my services with those being operated by kiddy-hosts and amateurs so I end up referring those customers to other budget providers who are setup to provide services for those client types. Because the prices are low doesn't mean that the service is inferior. It just means that you are trading a lot of non-essential frills and reliability for a lower price point. I can think of many multi-carrier, redundant this and that data centers for which customers pay a premium for only to end up suffering the same or if not more outages with those providers. For a catalog of this I will provide two references: [1] http://www.datacenterknowledge.com/archives/category/downtime/
and [2] http://www.webhostingtalk.com/forumdisplay.php?f=59
Therefore, for the price you are paying or the money you are saving , you should make appropriate investments to create your own redundency or tolerate the failures when they occur. (3) Due diligence: Because the main focus of my business is providing a platform for people to host applications and services. I have a lot of my money invested into this venture therefore, it is important to me to make that platform as reliable as possible within reason relative to the revenues generated by the service. Because increased system reliability equals increased customer confidence. Therefore, I am very proactive about keeping disruptions to a minimum. In fact, I think that you will find that this service has better than expected reliability for the prices.

Bandwidth Management
(1) The backbone bandwidth is shared by all service subscribers and is oversold by a standard ratio however, this does not mean that it is over-utilized. (2) Due to the bandwidth allocation policies, and default traffic control policies, any given subset of subscribers will be able to use 100% of the bandwidth allocated out of the total pool assigned to the servers as a group during the subscription period. For example, LWC10x5 comes with 1500GB/month of bandwidth. How you use this bandwidth is up to you but your average utilization should not be more than 5Mbps at any given time. If you are load-balancing traffic between all of your servers then each server has a volume transfer rate of 150GB per box. If you have an LWC20x5 plan you are splitting 1500GB over 20 machines now so each server has a volume transfer amount of 75GB per box. (2) Under the current bandwidth management policies, Each connection can burst up to 50% higher than the rated bandwidth allocation. Therefore, if you have and LWC10x5. All of your servers are connected to 100Mbit/s ports but your burst rate is set to 10Mbit/s. Therefore, as an aggregate your machines will be able to handle traffic spikes. (3) Special attention is paid to connections that have multi-megabit sustained usage and adjustments are made accordingly. Therefore, if you want dedicated bandwidth, you should be willing to pay dedicated prices. (4) Dedicated bandwidth is available for those willing to pay pass-through pricing. However, you should be aware that the cost of bandwidth plus the loop is going to be in $25~$30/Mbps range.

Bandwidth Management

(1) The backbone bandwidth is shared by all service subscribers and is oversold by a standard ratio however, this does not mean that it is over-utilized. (2) Due to the bandwidth allocation policies, and default traffic control policies, any given subset of subscribers will be able to use 100% of the bandwidth allocated out of the total pool assigned to the servers as a group during the subscription period. For example, LWC10x5 comes with 1500GB/month of bandwidth. How you use this bandwidth is up to you but your average utilization should not be more than 5Mbps at any given time. If you are load-balancing traffic between all of your servers then each server has a volume transfer rate of 150GB per box. If you have an LWC20x5 plan you are splitting 1500GB over 20 machines now so each server has a volume transfer amount of 75GB per box. (2) Under the current bandwidth management policies, Each connection can burst up to 50% higher than the rated bandwidth allocation. Therefore, if you have and LWC10x5. All of your servers are connected to 100Mbit/s ports but your burst rate is set to 10Mbit/s. Therefore, as an aggregate your machines will be able to handle traffic spikes. (3) Special attention is paid to connections that have multi-megabit sustained usage and adjustments are made accordingly. Therefore, if you want dedicated bandwidth, you should be willing to pay dedicated prices. (4) Dedicated bandwidth is available for those willing to pay pass-through pricing. However, you should be aware that the cost of bandwidth plus the loop is going to be in $25~$30/Mbps range.

Bandwidth Throughput / Performance
First, let me explain why the number of Servers on a segment is not a reliable metric for determining performance. It assumes that all of the Servers on the segment are busy accessing the network with high-volumes of traffic when in fact, the vast majority of Servers on a network segment(s) are idle or using very little traffic. Sure, there are some exceptions but this isn't the norm.

Bandwidth Abuse / Overages / Caps
(1) While I don't mind if people go over the monthly allocation within reason, the determination of how much of an overage constitutes abuse is at my sole discretion. (2) I will not hesitate to cap abusers. i.e. servers operating at double-digit bandwidth rates for hours on end. (3) The past-through cost to provide dedicated connectivity to any of my facilities is $25 per megabit that price does not include servers or power! Therefore, I will rate-limit, cap, throttle or bounce from the service those that are using bandwidth in a way that is costly or otherwise irresponsible. (4) I have no problem provisioning dedicated bandwidth for customers who are willing to pay the costs i.e. 10Mbit/s Unmetered connection (port + loop) $250/month.

Bandwidth Abuse / Overages / Caps

(1) While I don't mind if people go over the monthly allocation within reason, the determination of how much of an overage constitutes abuse is at my sole discretion. (2) I will not hesitate to cap abusers. i.e. servers operating at double-digit bandwidth rates for hours on end. (3) The past-through cost to provide dedicated connectivity to any of my facilities is $25 per megabit that price does not include servers or power! Therefore, I will rate-limit, cap, throttle or bounce from the service those that are using bandwidth in a way that is costly or otherwise irresponsible. (4) I have no problem provisioning dedicated bandwidth for customers who are willing to pay the costs i.e. 10Mbit/s Unmetered connection (port + loop) $250/month.

Control Panels
The servers do not come with control panels. There are several open source control panels available and I highly recommend that you use one that you have experience with installing.

Supported OS Linux Distribution
Let me be clear about what I mean when I say supported, it means that the hardware platform will run the Operating System as installed. The hardware environment is 64-bit Therefore, the latest 64-bit (x86_64) Linux Operating Systems are supported. CentOS and Ubuntu. By default, all servers are pre-configured with CentOS 5.x. If you want Ubuntu x64 please email sales@leeware.com prior to purchase please be aware that this will increase the deployment time as this is considered a custom install.

Supported OS Linux Distribution

Let me be clear about what I mean when I say supported, it means that the hardware platform will run the Operating System as installed. The hardware environment is 64-bit Therefore, the latest 64-bit (x86_64) Linux Operating Systems are supported. CentOS and Ubuntu. By default, all servers are pre-configured with CentOS 5.x. If you want Ubuntu x64 please email sales@leeware.com prior to purchase please be aware that this will increase the deployment time as this is considered a custom install.

IP-KVM and Out of Band Access
Another question I'm asked often, is if I provide any type of management console for people to restart, or reinstall the OSes on the Server? The short answer is NO. The longer answer is, those solutions cost money and will increase the costs significantly. Therefore, if you lose access to your machine you will have to send an email to admin to have it restored or rebooted (normal ETAs apply) so that you can get access to it again. Now, while this may seem like a pain for a provider it actually helps to cut down on carelessness. The incidents of lockouts and poor configurations are kept to a minimal. I think when there are real consequences for getting locked out of a server people naturally become experts at working on remote machines. Think before you type. There is a $25 charge for OS reloads and IP-KVM is available for short-term use which also requires a $25 charge. Both charges require pre-payment and the links for this is: http://www.leeware.com/rhands.html

IP-KVM and Out of Band Access

Another question I'm asked often, is if I provide any type of management console for people to restart, or reinstall the OSes on the Server? The short answer is NO. The longer answer is, those solutions cost money and will increase the costs significantly. Therefore, if you lose access to your machine you will have to send an email to admin to have it restored or rebooted (normal ETAs apply) so that you can get access to it again. Now, while this may seem like a pain for a provider it actually helps to cut down on carelessness. The incidents of lockouts and poor configurations are kept to a minimal. I think when there are real consequences for getting locked out of a server people naturally become experts at working on remote machines. Think before you type. There is a $25 charge for OS reloads and IP-KVM is available for short-term use which also requires a $25 charge. Both charges require pre-payment and the links for this is: http://www.leeware.com/rhands.html

Unmanaged Service Overview In Simple Terms
(1) By default all machines are patched. However, you are responsible for patching after you become the owner of the server. (2) The servers are already setup to a degree for basic things. That is basic tools are available such as editors and yum. (3) The firewall is turned off by default you have to turn it on if you want it. (4) Additional IP addresses can be rented for $1.50 each (http://www.leeware.com/rentxip.html) (5) Servers are located in un-staffed privately held Lights Out Facilities that are located throughout the Chicago Metro Area. The POP is located at 427 S. LaSalle (Cogent Communications formally PSI Net) If you need "premium bandwidth" be prepared to pay a "premium price". I can refer you out to an associate who has space @ 350 Cermak in Equinix. (6) With this service you get the following: (a) - A box with an Internet connection (b) The machine has an OS installed that has been patched and a running ssh server. (The firewall is turned off.) This setup is provided so that you can have the most control over your configuration. If you turn the firewall on (can be done from the command line) You will only be able to access the Webserver, ssh and nothing else unless you create a rule to allow it. You will lose access to the machine. (c) - If you need a DNS, FTP or other server (MAIL SERVERS won't work until you request a policy update from admin!). You should know how to install and configure those things. (remember google is your friend.) (d) - I TAKE A HANDS OFF APPROACH TO MACHINE MANAGEMENT Outside of administrative actions. (reboots and reloads) At sign-up you will be assigned a good password for your account however, it is always a good idea to change this password. I suggest that you use good password practices when selecting a password for your box. (e) - If you jack up the machine really bad. I won't waste anytime trying to figure out what you've done to it. I will just blow it away and give you a fresh one. (see fee schedule for reinstall charges.) (f) - The machines I provide can be used to learn how to manage a dedicated server or get started on a project while keeping your costs low. When you are ready for Enterprise Class Service, you can move on to another provider. (g) - Finally, it is always good to remind those who would SIGN-UP and after three days of fooling around with the machine demand their money back because they don't know how to install and FTP server. You will NOT BE REFUNDED YOUR MONEY! The aforementioned information has been provided to help the dedicated server shopper make an informed decision.

Unmanaged Service Overview In Simple Terms

(1) By default all machines are patched. However, you are responsible for patching after you become the owner of the server. (2) The servers are already setup to a degree for basic things. That is basic tools are available such as editors and yum. (3) The firewall is turned off by default you have to turn it on if you want it. (4) Additional IP addresses can be rented for $1.50 each (http://www.leeware.com/rentxip.html) (5) Servers are located in un-staffed privately held Lights Out Facilities that are located throughout the Chicago Metro Area. The POP is located at 427 S. LaSalle (Cogent Communications formally PSI Net) If you need "premium bandwidth" be prepared to pay a "premium price". I can refer you out to an associate who has space @ 350 Cermak in Equinix. (6) With this service you get the following: (a) - A box with an Internet connection (b) The machine has an OS installed that has been patched and a running ssh server. (The firewall is turned off.) This setup is provided so that you can have the most control over your configuration. If you turn the firewall on (can be done from the command line) You will only be able to access the Webserver, ssh and nothing else unless you create a rule to allow it. You will lose access to the machine. (c) - If you need a DNS, FTP or other server (MAIL SERVERS won't work until you request a policy update from admin!). You should know how to install and configure those things. (remember google is your friend.) (d) - I TAKE A HANDS OFF APPROACH TO MACHINE MANAGEMENT Outside of administrative actions. (reboots and reloads) At sign-up you will be assigned a good password for your account however, it is always a good idea to change this password. I suggest that you use good password practices when selecting a password for your box. (e) - If you jack up the machine really bad. I won't waste anytime trying to figure out what you've done to it. I will just blow it away and give you a fresh one. (see fee schedule for reinstall charges.) (f) - The machines I provide can be used to learn how to manage a dedicated server or get started on a project while keeping your costs low. When you are ready for Enterprise Class Service, you can move on to another provider. (g) - Finally, it is always good to remind those who would SIGN-UP and after three days of fooling around with the machine demand their money back because they don't know how to install and FTP server. You will NOT BE REFUNDED YOUR MONEY! The aforementioned information has been provided to help the dedicated server shopper make an informed decision.

Community Support Forums
(1) My customer-base is largely self-sufficient. Because I run a relatively large infrastructure I communicate problems and issues to Individuals as problems and issues are likely to only impact certain users and will remain irrelevant to all others. Therefore, If you need a certain level of comfort i.e. 24/7 support you should probably select a provider that offers that. (2) It's not a secret that am a solo operator. However, the nature of my business and my work requires that I have access to a computer(s) for up to 10-16 hours a day. With that being said, as per the TOS/AUP, you will generally get a response within 24 hours. Realistically, anywhere between 15 min - 4 hours. In other words, it could be more but usually less. (It should be obvious that I need time to, sleep, eat, exercise, have dinner with my wife, go on vacations, do research and development, setup new servers, work in the data center, solve problems, promote the business, do accounting and take a mental health break.) Also, because my customer base is spread out across the globe, non-US customers need to be mindful of the fact that while it may be in the middle of the day where you are, the opposite is true here so, I might be in bed and won't see your emails until I start the next day.

Community Support Forums

(1) My customer-base is largely self-sufficient. Because I run a relatively large infrastructure I communicate problems and issues to Individuals as problems and issues are likely to only impact certain users and will remain irrelevant to all others. Therefore, If you need a certain level of comfort i.e. 24/7 support you should probably select a provider that offers that. (2) It's not a secret that am a solo operator. However, the nature of my business and my work requires that I have access to a computer(s) for up to 10-16 hours a day. With that being said, as per the TOS/AUP, you will generally get a response within 24 hours. Realistically, anywhere between 15 min - 4 hours. In other words, it could be more but usually less. (It should be obvious that I need time to, sleep, eat, exercise, have dinner with my wife, go on vacations, do research and development, setup new servers, work in the data center, solve problems, promote the business, do accounting and take a mental health break.) Also, because my customer base is spread out across the globe, non-US customers need to be mindful of the fact that while it may be in the middle of the day where you are, the opposite is true here so, I might be in bed and won't see your emails until I start the next day.

5 Years Service Reliability / Redundancy SLA (99.96%)
I track two types of failures (a) Global Outages due to power / network issues which result in all customer systems being unavailable and (b) Equipment failures which result in outages for customers. (1) Global Reliability: In 5 years of operations and tracking (running for a lot longer) this service had been operating as of this posting for 42,700 hours. The service has an independently tracked uptime of 99.96%. The service has experienced a total downtime during that 5 year period of 18 hours with virtually all of the multi-hour outages being attributable to sub-customer activity. During that period, the various facilities have experienced up to 20 hours of power outages but due to UPS and manual switch generators, the customer impact of those outages are limited to just 2 hours and 36 minutes. (2) Network Carrier Reliability: In 5 years of operation only 2 hours of downtime can be associated with carrier activity. 45 minutes maintenance related and 10 minutes fiber-ring failure. (3) Equipment Reliability: With an aggregate of over 4TB of RAM in production RAM module failures are the most frequent cause of system failures. Followed by network cards/motherboards and then power supplies, cpu-fans and finally hard disk drives. (2) Independently maintained service statistics can be found here: http://www.webhostingstuff.com/uptime/LeeWare.html (3) Network SLA: (99.9%) Network Outages are beyond my control. Performance issues related to peering between ISPs are beyond my control. However, if I have any information about a planned outage or other issue that might cause problems I will post a notice on the status page http://www.leeware.com/status.html (4) Network Performance: The network features 99.9% availability with an average US latency of < 45.0ms New York to London < 85.0ms (5) Hardware SLA: (99.9%) If any hardware component fails I maintain enough spare equipment to replace the spare component or the entire machine within 24 hours. #4 Power SLA: (99.9%) Primary power is supplied by a public utility (all facilities are on the same power grid) the service is generally reliable. All services are protected via UPS against spikes, sags and short-term power outages. In the event of an extended outage, secondary power will be supplied by generator. (6) The battery backups are configured to run until dead therefore, a service outage will result when battery capacity is extremely low and the batteries drop the attached loads. When this happens, it can take 30-60 minutes to get all of the servers up on the generator partially due to a controlled startup protocol. All machines are set to stay down and not boot to avoid flapping, main breaker overloads or other overheat conditions due to the HVACs being offline. During the first quarter of 2011. Automatic Transfer Switches and Auto Start Generators will be installed at each of the locations. (7) Backups: Customers should backup their own data! (8) In a nutshell, I have customers from every corner of the planet and they try not to bother me too much and I try not to bother them unnecessarily. I'm not trying to convince you to use my service. I think the product pretty much sells itself. However, I am trying to give you information that could help you to make an informed decision which includes, managing risks so that you can pick a provider that suits your needs even if it is not my service but one of my competitors. The goal here is to be up front about the service so that you can set appropriate expectations.

5 Years Service Reliability / Redundancy SLA (99.96%)

I track two types of failures (a) Global Outages due to power / network issues which result in all customer systems being unavailable and (b) Equipment failures which result in outages for customers.

(1) Global Reliability: In 5 years of operations and tracking (running for a lot longer) this service had been operating as of this posting for 42,700 hours. The service has an independently tracked uptime of 99.96%. The service has experienced a total downtime during that 5 year period of 18 hours with virtually all of the multi-hour outages being attributable to sub-customer activity. During that period, the various facilities have experienced up to 20 hours of power outages but due to UPS and manual switch generators, the customer impact of those outages are limited to just 2 hours and 36 minutes. (2) Network Carrier Reliability: In 5 years of operation only 2 hours of downtime can be associated with carrier activity. 45 minutes maintenance related and 10 minutes fiber-ring failure. (3) Equipment Reliability: With an aggregate of over 4TB of RAM in production RAM module failures are the most frequent cause of system failures. Followed by network cards/motherboards and then power supplies, cpu-fans and finally hard disk drives.

(2) Independently maintained service statistics can be found here: http://www.webhostingstuff.com/uptime/LeeWare.html

(3) Network SLA: (99.9%) Network Outages are beyond my control. Performance issues related to peering between ISPs are beyond my control. However, if I have any information about a planned outage or other issue that might cause problems I will post a notice on the status page http://www.leeware.com/status.html

(4) Network Performance: The network features 99.9% availability with an average US latency of < 45.0ms New York to London < 85.0ms

(5) Hardware SLA: (99.9%) If any hardware component fails I maintain enough spare equipment to replace the spare component or the entire machine within 24 hours. #4 Power SLA: (99.9%) Primary power is supplied by a public utility (all facilities are on the same power grid) the service is generally reliable. All services are protected via UPS against spikes, sags and short-term power outages. In the event of an extended outage, secondary power will be supplied by generator.

(6) The battery backups are configured to run until dead therefore, a service outage will result when battery capacity is extremely low and the batteries drop the attached loads. When this happens, it can take 30-60 minutes to get all of the servers up on the generator partially due to a controlled startup protocol. All machines are set to stay down and not boot to avoid flapping, main breaker overloads or other overheat conditions due to the HVACs being offline. During the first quarter of 2011. Automatic Transfer Switches and Auto Start Generators will be installed at each of the locations.

(7) Backups: Customers should backup their own data!

(8) In a nutshell, I have customers from every corner of the planet and they try not to bother me too much and I try not to bother them unnecessarily. I'm not trying to convince you to use my service. I think the product pretty much sells itself. However, I am trying to give you information that could help you to make an informed decision which includes, managing risks so that you can pick a provider that suits your needs even if it is not my service but one of my competitors. The goal here is to be up front about the service so that you can set appropriate expectations.

Running IRC is Prohibited!
Running IRC is prohibited I don't care if it is a client or a server. I don't want it on my network. You risk server disconnection by breaking this rule.

Bittorrent & File Sharing and P2P is Prohibited!
Due to the way these protocols use bandwidth and the unwillingness of the operators of such applications to pay their share of the bandwidth bills, these applications are not allowed on the network. Furthermore, if such activity is detected your connection will be rate-limited and then dropped from the network.

Cancellation Policy Explained
Sometimes I get email from users who are frustrated that their servers were disconnected. These disconnections are usually the result of a subscription cancellation or payment failure. For security and management reasons servers are disconnected when cancellations are processed. (sign-up today cancel tomorrow) your server will be disconnected. Server payments are non-refundable. This is indicated in the very first line of the AUP/TOS as well as in the Welcome letter.

Cancellation Policy Explained

Sometimes I get email from users who are frustrated that their servers were disconnected. These disconnections are usually the result of a subscription cancellation or payment failure. For security and management reasons servers are disconnected when cancellations are processed. (sign-up today cancel tomorrow) your server will be disconnected. Server payments are non-refundable. This is indicated in the very first line of the AUP/TOS as well as in the Welcome letter.

NRC and MRC Explained
NRC (Non Recurring Charge i.e. one time payment) vs and MRC (Monthly Recurring Charge i.e. perpetual payments)

The High Cost of Low Prices
It is very important that you understand the risks associated with using this service. First, it is 100% no frills unmanaged service that is primarily intended for people who have experience with remote machine administration. All servers are housed in un staffed facilities where issues are addressed once or twice every 24 hours. Or when I can make the rounds to visit a particular location. Therefore, it is highly recommended that you not engage in activities that could keep your server from booting: In other words, try to avoid newbie errors such as; excessively rebooting your server after every change, implementing creative and untested start up configurations, screwing around with the system firewall (especially if you have no idea what you are doing!) installing software that trashes your system etc. Servers that have hung or crashed should be reported using the appropriate procedure for reporting such issues. These issues will be addressed during the daily server check. OS installs are a line-item and can be purchased for a fee. Or they can be done for free if done during a maintenance window up to 12-24 hours after the request.

The High Cost of Low Prices

It is very important that you understand the risks associated with using this service. First, it is 100% no frills unmanaged service that is primarily intended for people who have experience with remote machine administration. All servers are housed in un staffed facilities where issues are addressed once or twice every 24 hours. Or when I can make the rounds to visit a particular location. Therefore, it is highly recommended that you not engage in activities that could keep your server from booting: In other words, try to avoid newbie errors such as; excessively rebooting your server after every change, implementing creative and untested start up configurations, screwing around with the system firewall (especially if you have no idea what you are doing!) installing software that trashes your system etc. Servers that have hung or crashed should be reported using the appropriate procedure for reporting such issues. These issues will be addressed during the daily server check. OS installs are a line-item and can be purchased for a fee. Or they can be done for free if done during a maintenance window up to 12-24 hours after the request.

4GB RAM Installed but only 3.2GB Available?
While there is 4GB of RAM installed in every server. On motherboards that support a max memory configuration of 4GB 700~900MB of memory is used for addressing hardware devices That's why you will see a number greater than 3GB but less than 4GB. This is not a problem that can be addressed through software such as, kernel patches and 64-bit operating systems. For more information see: http://www.codinghorror.com/blog/archives/000811.html

4GB RAM Installed but only 3.2GB Available?

While there is 4GB of RAM installed in every server. On motherboards that support a max memory configuration of 4GB 700~900MB of memory is used for addressing hardware devices That's why you will see a number greater than 3GB but less than 4GB. This is not a problem that can be addressed through software such as, kernel patches and 64-bit operating systems. For more information see: http://www.codinghorror.com/blog/archives/000811.html

Renting a server from us is easy, as we accept PAYPAL and all major credit cards. Typical setup times are less than 24 hours.
Get your Linux Server today!

Home

About Us

Products

Contact Us

About LeeWare

Products

Related Links

Company

Analysis