Distribution One Server Requirements
Automation
Experience
Technology
Connectivity
Flexibility
Support
Distribution One Server Requirements
Introduction
Welcome to the Hardware Configuration Guide. The goal of this guide is to provide a practical approach to sizing your
Distribution One application and database server to obtain the best performance possible from your Distribution One
software. It will also help you plan for the future growth of your system. The key to success in getting your hardware sized
correctly is to define your application load correctly and then to match it with the appropriate hardware resources. Choosing a
hardware architecture which can grow as your business and / or application load grows is also important. It is difficult to
accurately provide a hardware recommendation that matches your unique application usage pattern. Many customers find
this guide very useful and accurate after they have followed all the steps described in the sections. There are some cases
where the provided recommendations may not fully work out in your unique situation. It is therefore difficult to guarantee or
imply that the hardware purchased using this hardware sizing guide will solve your unique business requirements. To improve
the odds of success, we highly recommend that you work with the Distribution One Technical Support group and implement
their hardware recommendations.
Virtual Deployment of Distribution One Applications. Virtualization is an abstraction layer that decouples the physical
hardware from the operating system to deliver enhanced IT resource utilization and flexibility. There are various software
products available that can be used to build your virtual deployment; Hyper-V or VM-Ware are the most popular.
Virtualization is great for disaster recovery and quickly creating new servers. When designing the virtual server infrastructure,
please ensure that you are matching the virtual machine’s specifications with the hardware specifications mentioned in this
guide. When matching hardware specifications with your virtual server machine you must make sure you are not
interchanging sockets with cores. Each virtual machine instance consumes anywhere from 5-15% per virtual machine on the
host machine, depending on the host machine’s hardware. Please take that additional usage into consideration as well.
Typical Usage. The hardware recommendation provided in this guide should work for a wide variety of customers. But it is
also normal to expect that typical load assumes that during the very busy times of the day or month the load on the server will
be higher than usual. We expect the server’s response to be slower than less busy times of the day. Unfortunately, we cannot
anticipate what the peak load for your business practice / application usage will be. If you find the server response
unacceptable, upgrade the appropriate component which is under pressure to remedy the situation. It is important to choose
a hardware architecture which allows for expandability. Distribution One Technical Support can also help you here.
Performance Tuning. Relational databases are advanced systems that sometimes require tuning for optimal performance.
One major selling point of both platforms is that the database engines are largely self-tuning. However, it can still be useful to
perform some level of tuning on a database particularly in the area of database input / output operations to the physical disk
subsystem. If performed by an experienced professional, this tuning can substantially improve performance on your
hardware. Note that there are many considerations, outside of the scope of this guide, that are involved in maintaining the
server at peak performance.
Don’t run other Applications on the application and database Server. Due to the mission critical nature of any company’s
Enterprise Resource Planning (ERP) system, it is recommended that you have a server dedicated to running your Distribution
One production instance. Running other applications including Test / Training / Pilot / Development implementation of the
Distribution One application on the database and / or separate application server can degrade performance. While certain
situations may be suitable for dual hosting, it is strongly recommended that Distribution One run on a dedicated machine.
Microsoft Small Business Server (SBS) is not a recommended platform for Distribution One due to the multi-application design
of that product.
Support: (856) 380-0626
www.distone.com
Page 1
Proprietary Information;
Do Not Distribute
Automation
Experience
Technology
Connectivity
Flexibility
Support
Direct Network Connection Between the Application Server and Database Server. When you separate the database server
from the Distribution One Application Server, it has to be directly connected with at least a 1 Gigabit network connection
between the application and database servers. If this condition is not met the application performance will suffer. We don’t
want any latency between them.
Disaster Recovery and High Availability. Unfortunately it is beyond the scope of this guide to cover the details of this topic.
Distribution One Technical Support group is a great resource for this discussion and they will be happy to work with you to
make recommendations.
Sample Server Configuration 5-15 users
•
•
•
•
•
•
•
•
Virtualization (recommended)
Quad-core processor
12 GB RAM – the more the better
IOPS 10,000+
500GB, SSD Drives with RAID10 (recommended)
Separate ARRAY or controller for the OS partition and database partition
“E:” Drive Partition for the Database
Internal or External Modem – For Outbound Faxing
Sample Server Configuration 16-30 users (2 servers)
•
•
•
•
•
•
•
•
Virtualization
Quad-core processor
12 GB RAM – the more the better
IOPS 50,000+
500GB, SSD Drives with RAID10
Separate ARRAY or controller for the OS partition and database partition
“E:” Drive Partition for the Database
Internal or External Modem – For Outbound Faxing
Sample Server Configuration 31-50 users (2 servers)
•
•
•
•
•
•
•
•
Virtualization
Quad-core processor
32 GB RAM – the more the better
IOPS 50,000+
500GB, SSD Drives with RAID10
Separate ARRAY or controller for the OS partition and database partition
“E:” Drive Partition for the Database
Internal or External Modem – For Outbound Faxing
50+ Users Contact Distribution One Technical Support
Required Software
• Windows Server 2012 Standard or Enterprise Operating System (64-bit)
• Windows Server CAL’s (1 per User)
• Windows Remote Desktop Services CAL’s (1 per User)
• Microsoft Word & Excel or Office 2003 or later – This is for the application only not at the PC level.
Support: (856) 380-0626
www.distone.com
Page 2
Proprietary Information;
Do Not Distribute
Automation
Experience
Technology
Connectivity
Flexibility
Support
Backup, Fault Tolerance UPS, Security Software
Backup Devices
It is extremely important that your business information is backed up on a regular basis. The frequency and type of backup
should be discussed with your implementation team, but one or more tape / backup drives on the server can ensure a timely
backup is made. It is also possible to back up to the local disk, and then copy the file elsewhere for storage. Consider off-site
storage for backups as well. Test your backup scripts and restore procedures and backup files regularly.
Uninterruptible Power Supply (UPS) and Redundant Components
The production server is the key to the business, and should be protected. Power outages can damage the server hardware
and can occasionally corrupt the software; therefore, the server should be protected with a UPS, specifically one with server
software that is capable of initiating a clean shut-down of the server. Recovery from a power outage is usually not a problem,
but why take the risk? The mission critical nature of an ERP system necessitates consideration of redundant processors,
network controllers, storage controllers, hot-swappable hard drives, and power supplies. You should also consider network
hub / switch redundancy. Such components contribute to a “highly - available” solution which is tolerant of most hardware
failures and allow for service to be postponed to off-hours. Businesses requiring continuous availability (24 x 7) may also want
to consider hot plug support for controllers, on-line spare hard drives, redundant processor power modules, and redundant
hot-swappable fans to provide the highest-availability solution that can be serviced during business hours without an
interruption in productivity. Network hubs, routers, and server monitors should also be considered when assessing UPS
requirements.
Security Software
The same sorts of considerations you have for your desktop anti-malware system are also key for your server-based products.
Significant features to look for include the ability to approve or decline signature updates centrally, the ability to initiate scans
on remote machines from a central administrative console, logging and reporting functionality that is accessible both from a
console and over the Web, and the ability to issue e-mail or text message alerts upon detection of suspicious activity or a
confirmed outbreak. Particularly for your servers, you need some sort of real-time monitoring capability so you can see all of
your servers, and their malware protection status, at a glance. (Some desktop anti-malware products do not offer this.)
The bottom line is that it is as important as ever to keep a watchful eye for malware on all levels of your IT infrastructure.
Keeping your desktops free of malware is a great goal, but keeping your servers free from distributing the malicious software
both internally and externally is an essential policy to make and uphold.
Disk System – Sizing, Configuration and RAID
Sizing Disk Subsystem
There are three primary areas to consider when sizing the disk sub system.
• Capacity
• Availability
• Performance – RPM and multiple spindles writing at once in a RAID configuration to increase the read / write
performance. Consider each of these factors jointly when configuring your I / O system.
How Much Disk Capacity?
Capacity is the initial concern for the disk subsystem, but too often its relationship to availability and performance are not
taken into consideration. Capacity is the easiest resource to calculate. The size of your Distribution One database will depend
on the number of users and their transaction volume. In general, Distribution One databases range from 300 MB for smaller
systems to over 200 GB for larger systems. Consider that many customers use training databases in addition to a production
database. You may also have the requirement to evaluate a new version of Distribution One in a test environment prior to
upgrading your production database.
Support: (856) 380-0626
www.distone.com
Page 3
Proprietary Information;
Do Not Distribute
Automation
Experience
Technology
Connectivity
Flexibility
Support
What Types of Drives Should I Purchase?
Always select the fastest drives and disk adapter technology available within your budget. For best performance, select SCSI
Ultra-Fast / Wide Disk Adapters, Ultra-Fast / Wide Disk Drives or Serial Attached SCSI (SAS) rotating at 15,000 rpm. Select a
higher value of disk buffer size like 16 or 32 MB or more. Solid State SSDs are a great choice especially for larger systems or
greater performance.
How Do I Protect My Data?
Availability is the concept that your data is important enough to require some level of protection while keeping the cost of the
server within your budget. Protecting your data is the most important consideration. There are many ways to protect your
data from potential disaster, which include backups, security software, and hardware protection. For this discussion we will
focus on hardware protection, and more specifically Redundant Array of Independent Disks (RAID). See the “Fault Tolerant
RAID Configurations” section for additional details.
READ and WRITE Cache
Systems that contain a large READ disk cache within the controller can dramatically improve the performance of the
application server. This option is relatively inexpensive when purchased with the overall system and therefore is
recommended, particularly for larger implementations. WRITE caches can cause integrity issues under extremely rare and
complex circumstances. Most caches can be configured easily enough to avoid this rare issue by turning off the write cache.
System Drive Configuration (Spread I / O Load to Gain Performance)
The number of drives you purchase and how you decide to configure them depends on your budget, the intensity of your
transactions, the desired performance and the need for fault tolerance. Your budget determines the number and speed of the
disks you can afford. The intensity of transactions predicts what disk I / O you will need. Remember that the disk size is often
not the critical factor, but rather the speed in writing to and reading from the drives. Therefore, more drives mean these
operations can be spread out over more disk heads. Implementing RAID will help reduce administration, increase
performance, and provide fault tolerance, depending on which RAID type is chosen.
The major components of the Distribution One system include:
• Operating system
• Application code
• Application generated reports
• Progress database
• Database bi files
• Database data
In an ideal system, each one of these items would have its own disk, with the database file being located on a RAID10 set and
log files being mirrored for maximum performance and integrity. There are always some tradeoffs between the number of
disks employed and performance and data integrity. The initial recommendation would be to never locate the database or log
files on anything but fault tolerate drives. Your data is worth far more than the additional cost of the required drives. The
components that might be combined are the operating system and the application code. The log file should be the last
component you combine with something else. Also, consider having a spare drive or two on the system. You will often find
you need them later. Spare drives will let you have a copy of the database for testing, provide space for expansion during a
version upgrade, or provide a formatted drive in case of disk failure. The drive configurations listed below illustrate the
tradeoff between cost and performance and fault tolerance for each database component listed above. Performance for the
listed configurations may be enhanced by adding additional drives to the RAID arrays. Base your drive capacity decisions on
estimated future growth / incremental cost.
Growth Path. Buy a system which allows you to put several disks internally. Remember, we add more disk spindles not to add
capacity but to spread the I / O load across multiple spindles.
Support: (856) 380-0626
www.distone.com
Page 4
Proprietary Information;
Do Not Distribute
Automation
Experience
Technology
Connectivity
Flexibility
Support
Fault Tolerant RAID Configurations
One of the most powerful methods of improving the performance of any drive subsystem is to employ multiple drives working
in tandem. This is especially important for the database server storage subsystem. The reason for this is simple: there will
always be limits to the performance that a single drive, no matter how fast, can achieve. This limit is derived from the fact that
one drive can only perform one operation at a time, having only one set of read / write heads. A multi-drive configuration can
deliver some amazing performance gains when used with an operating system that supports an asynchronous I / O model and
device concurrency (talking to multiple devices simultaneously) such as Windows Server 2008 or 2012 (x64). The most
common reason for using multiple drive sets is their use in RAID configurations. A RAID array is a collection of disks that
provide a single logical volume to the operating system and user, faster performance than a single logical drive by itself, and in
most implementations, some form of fault tolerance. For RAID disk arrays, fault tolerance means that one member disk in a
set can fail and the operating system or RAID hardware can recover and rebuild the information from the data contained on
the other member drives in the set. RAID is defined in a series of levels (starting with level 0), most of which provide some
form of data fault tolerance in addition to their speed advantages. Because the data redundancy and performance
implications of RAID are so significant, their use warrants consideration by anyone looking to achieve maximum disk
subsystem performance.
The following RAID configurations are supported with your Distribution One applications:
• RAID 0. Also known as disk striping. RAID 0 is the only common implementation of RAID that offers no fault tolerance.
It does, however, offer the fastest performance of any RAID level; it works by striping data evenly across two or more
drives in the RAID 0 set. This concurrent drive usage translates to excellent read/write performance and is the best
choice for maximum performance where data fault tolerance on the drive array itself is neither a concern nor a
requirement.
•
RAID 1. Also known as disk mirroring or duplexing. With RAID 1, two drives of equivalent capacity are used. One is a
“mirror” or “shadow” of the other. Disk mirroring provides a redundant, identical copy of a selected disk - all data
written to the primary disk is also written to the mirror disk. RAID 1 provides the best fault tolerance because no
rebuilding is required in the event of a failure. RAID 1 generally improves read performance but may degrade write
performance. When two controllers are used in RAID 1 (one for each drive), this is referred to as disk duplexing.
• RAID 10 (RAID 1 + 0). Also known as mirroring with striping. RAID10 uses a striped array of disks, which are then
mirrored to another identical set of striped disks. For example, a striped array can be created using three disks. The
striped array of disks is then mirrored using another set of three striped disks. RAID 10 provides the performance
benefits of disk striping with the disk redundancy of mirroring, without the performance issues associated with parity
maintenance. RAID 10 provides the highest read/write performance of any of the RAID levels at the expense of using
twice as many disks. This is advised for high security / high performance situations. Distribution One recommends
RAID 10 for database files and Distribution One application server’s temporary and report files.
• *****RAID 5 (or any other combination similar to RAID 5).***** Not recommended and supported. It does not
mean that Distribution One applications will not work. It means that performance will suffer considerably if you
configure your disks in a RAID 5 or any other RAID 5 combination.
Support: (856) 380-0626
www.distone.com
Page 5
Proprietary Information;
Do Not Distribute
Automation
Experience
Technology
Connectivity
Flexibility
Support
Implementation Types
The four common RAID implementations include:
• In the operating system
• In the disk controller
• Via a dedicated RAID controller hardware
• In a separate subsystem
Implementing RAID as part of the operating system is the least expensive option and also the least desirable. Software
implemented RAID solutions lack hardware level control and place additional load on the server, reducing the amount of
server resources available for your Distribution One applications.
Sizing Memory Requirements (RAM)
This section is provided to give you an idea of how we calculate how much RAM is required. The RAM we have specified in the
hardware sizing tables that matches your user count and application load profile supersedes if it exceeds the calculation you
determine from this section. For a starting point, be sure you take into consideration the following:
• All applications that the application server and / or database server is running
• Windows 2012 x64 Server requirements
• Windows 2012 x64 Server File System Cache use
• The number of concurrent Distribution One users supported by the system
• The amount of server resources configured (CPU, Disk, Network, etc.)
• Performance expectation - for most environments, larger amounts of RAM will yield higher levels of overall
performance
Sizing Processor the CPU(s)
The primary object when sizing the model and number of CPU’s is related directly to the type of workload, the number of
users (peak load) and the expected volume of transactions for your Distribution One system. These key variables determine
the overall demand that will be placed on the database server. The typical rule of thumb is to configure the fastest CPU that is
economically feasible. Business requirements and the resources needed to support server-based applications are constantly
increasing. Therefore, starting with the fastest possible CPU and a server architecture that can support multiple CPUs provides
better long-term investment protection. Windows Server 2008, 2012 x64 are both multi-process and multi-threaded enabled
operating systems. Without any real applications running under Windows, there are multiple active processes and threads.
Thus, additional CPUs are normally helpful to a degree. Also, consider that the Progress database is designed for a Symmetric
Multi-Processing (SMP) environment and that the Distribution One application is multi-threaded and will scale to multi
processors when present. Distribution One would recommend starting with at least a Quad-Core processor server with a large
amount of L3 cache. Configure enough CPU resources to drive all of the server’s resources. Develop the server configuration
such that the CPU is “fed.” This is achieved by ensuring there is enough CPU cache, main memory (RAM), and I / O channels
for the disk and network devices. Windows Server 2008, 2012 x64 and other servers can easily be slowed by its weakest
component in the data path that is not providing the data requested by the CPU.
For optimal performance, configure the CPUs in the following way:
• Enable Hyper-Threading
• Disable Power Saving settings including C-States (in BIOS)
• Enable Turbo Boost
Support: (856) 380-0626
www.distone.com
Page 6
Proprietary Information;
Do Not Distribute
Automation
Experience
Technology
Connectivity
Flexibility
Support
Growth Path. If you have empty CPU sockets on your server’s mother board, consider adding extra CPUs to satisfy demands
on your Distribution One installation put by concurrent users. Distribution One’s architecture allows for splitting the
application and database to separate servers. You can also add multiple application server machines to allow for your growth.
Communications
A broadband internet connection via a Remote Desktop Connection through a static IP is required to access the Distribution
One Online Support Center for interim and commercial releases and remote support. DSL or higher quality connection
recommended. Internet Explorer 7.0 or higher is needed for Distribution One support.
Optional Equipment
Handheld Equipment
Device Requirements
• Screen size minimum of 240 x 320
• Device must support MS Telnet Services client
• IEEE 802.11 Radio Frequency standards support
Access Points
• IEEE 802.11 Radio Frequency standards support
Barcode Equipment
Support for Bar 39 Barcode
Printer
At least one laser printer is required for standard reports and forms. Must have current Windows drivers
Backup Equipment
Distribution One recommends your company invests in a reliable backup system. Your hardware vendor can recommend and
install one for you.
Support: (856) 380-0626
www.distone.com
Page 7
Proprietary Information;
Do Not Distribute
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising