How Data Centers Can Prepare for a Natural Disaster

May 31, 2018 at 8:00 AM

 

We’ve learned that the Cloud isn’t actually floating in the sky, it’s actually thousands of data centers full of servers that store and transmit all of the data we use every day. But what happens when a data center is affected by a natural disaster? In this post, we’ll take a look at what defensive strategies are used to keep our data safe and our cloud aloft, even in the worst circumstances.

 

From blizzards and hurricanes to floods and fires, we seem to have seen a large number of natural disasters in recent history. Fortunately, data centers are prepared with plans to maintain Internet connections and keep your data accessible even in the worst conditions. By having preparedness plans in place, staff willing to stay at their posts and generators to provide power, key data centers can withstand record-breaking hurricanes and even serve as evacuation shelters for citizens and headquarters for law enforcement.

 

Here are some ways data centers can prepare for natural disasters:

 

Make a Plan

The best line of defense is having a good offense - having a plan in place, testing that plan, having a plan B for when that plan fails and then being ready to improvise. When it comes to Mother Nature, even the most prepared have to roll with the punches as things change.

 

Build a Fortress

The ideal structure to house your data center will be impenetrable. That might be too much to ask, but newly constructed buildings can be made to withstand earthquakes, flood, fire or explosion. The addition of shatterproof/explosive-proof glass, reinforced concrete walls and being in a strategic location outside flood zones can also provide an extra layer of protection.

 

These additional precautions might not be possible in older buildings, but there are still steps you can take to help protect your data center:

 

·       Move hardware to a safer location if possible:

    - Ideally, a data center should be away from windows, in the               center of a building and above ground level

    - Higher floors are better, except in an earthquake zone, then               lower floors are safer

·       Install pumps to remove water and generators to keep the pumps          running

·       If there are windows, remove objects that could become airborne

·       Fire extinguishing systems should be checked regularly

 

Redundancy is Key

Hosting all data in one place is opening the door for disaster. A safer option is to host it in multiple locations at redundant centers that can back each other up if disaster strikes one or more facilities. These centers don’t have to be on opposite ends of the world, but putting them in different geographic regions is probably the safest bet. They should be far enough apart that one disaster won’t take them all out.

 

Back That Data Up

If there’s no time to back up data to the Cloud, making a physical backup of the data and sending it with someone who’s evacuating is a good second option.

 

Natural disasters are unavoidable, and the most important asset to keep safe is always the people working inside the data center, but with a plan in place to keep Mother Nature at bay, you might be able to salvage the data center too.

 

IEEE-488 the Intrepid Interface

May 24, 2018 at 8:00 AM

 

Developed by Hewlett Packard in the 1960s, IEEE-488 was designed to easily interconnect controllers and instruments.  Originally named HP-IB (Hewlett Packard Interface Bus), and now largely known as GPIB (general purpose interface bus), the interface was adopted and renamed by various institutions and standards bodies including the IEEE, which named it IEEE-488. No matter what you call it, this die-hard interface has endured throughout the years and is still widely used.  

 

The basic concept of IEEE-488 is an extremely flexible system that allows data to be transmitted between any instrument on the bus. It has a 24-pin connector and is double-headed, both ends of the cable are used, with female on one end and male on the other. There are 16 signal lines, eight for bi-directional communication, five for bus management and three lines are used for handshakes.

 

The speed of the data flow is determined by the slowest active instrument and the max data rate is around 1 Mbps. As many as 15 instruments can be connected on a bus with a maximum total length of 20 meters and no more than 2 meters between two adjacent instruments. Active extenders can be used to create longer buses that can connect up to 31 devices.

 

The way the IEEE-488 operates is that each instrument on the bus has a unique address and no two instruments can have the same address. The connected equipment is allocated into three categories, with some items serving more than one function (i.e. listener/talker):

 

  • · Controller – The controller is in charge of controlling the operation of the bus. This device, usually a computer, instructs the other instruments to perform in their various roles. The controller also makes sure that there aren’t any conflicts on the bus. For example, if two talkers talk at the same time, data would be corrupted and the system operation would be compromised. Several controllers can share the same bus, but only be one can be active at a time.

 

  • · Listener – A listener is connected to and accepts instructions from the bus. One example would be a printer that only accepts information.

 

  • · Talker – A talker returns data and instructions to the bus. For example, a meter being used on the bus would be a talker when it’s providing data.

 

Today, IEEE-488 is the most common communication interface for scientific and engineering instruments, it is also used in a wide range of applications including data acquisition. Most bench instruments accept IEEE-488 as a standard fitting or as an option to make it easy to use test equipment in various applications, aside from only being used in an ATE test stack. Though the interface is not standard on today’s computers, a GPIB card can be easily installed. Because of its flexibility and convenience, IEEE-488 and is still widely used today and will likely remain relevant for years to come.

 

411 on M12 Connectors

May 17, 2018 at 8:00 AM

 

Since their introduction in 1985, M12 connectors have grown to become the go-to interconnect system for industrial automation. These rugged connectors provide reliable connections in the harshest environments and have revolutionized the world of industrial automation connectivity.

 

M12 connectors are circular connectors that have a 12-mm locking thread and often boast IP ratings for protection against liquids and solids. They are ideal for connecting sensors, actuators, as well as industrial Ethernet and Fieldbus devices, mostly in industrial automation applications and in corrosive environments.

 

Prior to the inception of the M12 connector, engineers had to hard wire or repeatedly replace connectors that couldn’t endure in harsh conditions. Initially released with 3 and 4-pin models, the original M12 connector had a smaller current than its predecessor, the RK30, but still provided the protection of an IP67 rating. The 4-pin M12 connector allowed a single system to include more advanced sensors and actuators. Today, these rugged connectors are available with 3, 4, 5, 8 and 12-pin configurations with additional locking styles continuously being developed, such as bayonet and push-pull.

 

In addition to factory automation, M12 connectors and M12 cable assemblies are used in measurement and control, communications, transportation, robotics, agriculture and alternative energy applications. Choosing the correct pin count depends on the specific application. Three and 4-pin models are needed for sensors and in power applications. Ethernet and PROFINET require 4 and 8 pins. DeviceNet and CANbus mostly use 4 and 5-pin connectors. Twelve-pin models are typically specified for various signal applications.

 

Along with different pin counts, M12 connectors have multiple styles of key coding to prevent improper mating.  Here are the most common coding types and what they’re used for:

 

·       A-coded: sensors, DC power and 1 Gigabit Ethernet

·       B-coded: PROFIBUS

·       C-coded: AC power

·       D-coded: 100 Mbit Ethernet

·       X-coded: 10 Gigabit Ethernet

·       S-coded: AC power (will be replacing C-coded power parts)

·       T-coded: DC power (will be replacing A-coded power parts)

 

The most popular types of M12 coding are A, B, D and X.  The A, B and D-coded connectors are some of the first M12 connectors and have been on the market the longest. X-coded connectors are rising in demand for high-speed industrial Ethernet and will ultimately take the place of A and D-coded parts in Ethernet applications. The newest code designs being developed are K-coded for AC power and L-coded for PROFINET DC power.

 

Case Study: Lineage Logistics

May 10, 2018 at 8:00 AM

 

As the second largest cold storage thrid-party provider in the US, Lineage Logistics provides cold storage solutions for leading grocery, food and retail  companies.

 

During a wireless network roll-out in their warehouses, Lineage Logistics needed to add multiple access points inside refrigeration units with temperatures as low as -40 degrees. The company’s engineers had designed a wireless access point that fulfilled the needs of the project, but wasn’t able to function unprotected in the cold temperatures.

 

Lineage Logistics began searching for an enclosure that would protect their access point, offer Power over Ethernet (PoE) over a single cable and meet cost requirements.

 

After rejecting a competitor’s offering, as it didn’t fully meet the requirements, Lineage Logistics came to L-com for help. Our team was able to develop a comprehensive solution that met all their needs. We created a custom NEMA enclosure that could house the access point and provide plently of room if adjustments were needed. We were also able to save Lineage Logistics time and money on installation by mounting the access points in the enclosures and providing all required cabling and antennas.

  

L-com was able to not only meet, but surpass this client’s expectations with the perfect solution to their problem.

 

To read the full case study, click here.

 

802.3bz Provides Congestion Relief – 2.5 Gbps & 5 Gbps Over Copper

May 3, 2018 at 8:00 AM

 

Cat5e and Cat6 cables are two of the most widely used cables in the world. Traditionally, for conventional Cat5e and Cat6 twisted-pair copper cabling, Gigabit Ethernet (1 Gbps) is the fastest standard. A wired connection of 1 Gbps is probably enough speed for one PC user, but with the surge of high-speed Wi-Fi devices being used over the last few years, Gigabit Ethernet has become increasingly congested. Thus, the IEEE has developed the 802.3bz standard to ease the pain of 1 Gbps traffic and allow speeds of up to 2.5 Gbps and 5 Gbps over Cat5e and Cat6 copper cables.

  

To escape the 1 Gbps bottleneck and increase speeds to 10 Gbps, a network cable upgrade to Cat6a or Cat7 is usually required. At an estimated $300 per cable pull, upgrading cable is a costly process and not always feasible, especially for large networks which could also encounter expensive delays and connection disruptions in the process.  Fortunately, the 802.3bz allows users to avoid expensive cable upgrades. This new 2.5G/5GBASE-T standard can provide 2.5 Gbps speeds over 100 meters of Cat5e cable and 5 Gbps speeds over 100 meters of Cat6 cable. These higher speeds are bookended by a switch on one end and either an Ethernet extender or electronic device on the opposite end.

 

The physical layer of 2.5G/5GBASE-T is similar to 10GBASE-T, but uses 200 MHz or 100 MHz spectral bandwidth instead of 400 MHz. This is beneficial because 2.5G/5GBASE-T consumes less than half the bandwidth of 10GBASE-T and doesn’t require a high-quality, mega-shielded cable. The 802.3bz standard also provides additional features such as Power over Ethernet (PoE), which is useful when rolling out Wi-Fi access points.

 

With a growing need for faster connections, 802.3bz provides a sensible way to upgrade networking capabilities without the expense of re-cabling, all while improving user experience and avoiding costly downtime.

 

© L-com, Inc. All Rights Reserved. L-com, Inc., 50 High Street, West Mill, Third Floor, Suite 30, North Andover, MA 01845