* Click to see larger version
Archive for the ‘Uncategorized’ Category
Solid-State Storage Meets the Data Center
Solid-state-storage is already found in USB memory sticks, smartphones, media players and a variety of other devices. Before long, you may also find solid-state drives (SSDs) performing key tasks inside your data center.
While hard drives aren’t yet ready to join the endangered species list, SSDs can help data centers keep pace with speedier servers and applications. Here’s what you need to know about this emerging data center technology:
What is Solid-State Storage? Also known as flash memory, solid-state storage is a non-volatile, removable storage medium that uses integrated circuits (ICs) rather than magnetic or optical media.
Why is It Needed? As server CPU speeds keep accelerating, hard drive access times are struggling—and increasingly failing—to keep pace. This bottleneck can be felt in a number of areas, including database and caching applications, particularly when they are offered as a service or hosted within a cloud computing environment. SSDs can bring speedier access times to data center storage systems that hard drive vendors and users can currently only dream of.
Enterprises may also want to adopt SSDs to take advantage of the technology’s other benefits, including low power consumption, low heat generation, non-volatility, enhanced reliability and compact form factor characteristics—all appealing attributes in a colocation or general data center environment.
How Much Does It Cost? This is perhaps the biggest question. It’s widely known that solid-state storage is far more expensive than magnetic storage technologies, although the price-gap is getting smaller over time. In-Stat, a semiconductor industry research firm based in Scottsdale, Ariz., reports that flash prices have dropped an average 60 percent per year over the past three years. Nevertheless, SSDs remain far more expensive than equivalent-capacity hard drives (and it’s important to remember that hard drive prices are falling dramatically, too). In a nutshell, the price per gigabyte ratio between SSDs and HDs currently stands at between 10:1 and 20:1, depending on device capacity and quality.
Who Offers It? Nearly all server vendors already offer SSDs in one or more of their systems, either as the primary storage device or as an add-on unit. Storage system vendors, meanwhile, are marketing various types of SSDs as direct replacements for existing server hard drives and for use in storage area networks (SANs). Some companies are also building SSD technology into data access appliances targeted at high-demand database customers.
Who Will Benefit Most From the Technology? Basically, any enterprise that needs to run applications requiring high-speed data location and/or retrieval, including credit card transactions, video on demand and search engines.
When Will We Be Able to Get Rid of Our Hard Drives? Not at any point in the foreseeable future. In fact, data center hard disk eradication is a task your successor (or his or her successor) may face. In-Stat forecasts that by 2013 SSDs, will account for only 5.4 million of all enterprise mass storage device shipments (compared to 42.7 million hard drives). Interestingly, Gartner, the Stamford, Conn.-based technology research firm, sees cloud-based storage as a more serious and immediate threat to data center hard drives than solid-state storage.
Over the long haul, however, time and technology favors SSDs. In-Stat projects that SSD shipments will grow at a compound annual growth rate of 154.8 percent between the years 2007 and 2013 (as opposed to just 4.8 percent for hard drives). With such disparate growth rates, it appears highly likely that SSD deployments will likely one day catch up with and overtake hard drive installations.
Solid-state-storage is already found in USB memory sticks, smartphones, media players and a variety of other devices. Before long, you may also find solid-state drives (SSDs) performing key tasks inside your data center.
While hard drives aren’t yet ready to join the endangered species list, SSDs can help data centers keep pace with speedier servers and applications. Here’s what you need to know about this emerging data center technology:
What is Solid-State Storage? Also known as flash memory, solid-state storage is a non-volatile, removable storage medium that uses integrated circuits (ICs) rather than magnetic or optical media.
Why is It Needed? As server CPU speeds keep accelerating, hard drive access times are struggling—and increasingly failing—to keep pace. This bottleneck can be felt in a number of areas, including database and caching applications, particularly when they are offered as a service or hosted within a cloud computing environment. SSDs can bring speedier access times to data center storage systems that hard drive vendors and users can currently only dream of.
Enterprises may also want to adopt SSDs to take advantage of the technology’s other benefits, including low power consumption, low heat generation, non-volatility, enhanced reliability and compact form factor characteristics—all appealing attributes in a colocation or general data center environment.
How Much Does It Cost? This is perhaps the biggest question. It’s widely known that solid-state storage is far more expensive than magnetic storage technologies, although the price-gap is getting smaller over time. In-Stat, a semiconductor industry research firm based in Scottsdale, Ariz., reports that flash prices have dropped an average 60 percent per year over the past three years. Nevertheless, SSDs remain far more expensive than equivalent-capacity hard drives (and it’s important to remember that hard drive prices are falling dramatically, too). In a nutshell, the price per gigabyte ratio between SSDs and HDs currently stands at between 10:1 and 20:1, depending on device capacity and quality.
Who Offers It? Nearly all server vendors already offer SSDs in one or more of their systems, either as the primary storage device or as an add-on unit. Storage system vendors, meanwhile, are marketing various types of SSDs as direct replacements for existing server hard drives and for use in storage area networks (SANs). Some companies are also building SSD technology into data access appliances targeted at high-demand database customers.
Who Will Benefit Most From the Technology? Basically, any enterprise that needs to run applications requiring high-speed data location and/or retrieval, including credit card transactions, video on demand and search engines.
When Will We Be Able to Get Rid of Our Hard Drives? Not at any point in the foreseeable future. In fact, data center hard disk eradication is a task your successor (or his or her successor) may face. In-Stat forecasts that by 2013 SSDs, will account for only 5.4 million of all enterprise mass storage device shipments (compared to 42.7 million hard drives). Interestingly, Gartner, the Stamford, Conn.-based technology research firm, sees cloud-based storage as a more serious and immediate threat to data center hard drives than solid-state storage.
Over the long haul, however, time and technology favors SSDs. In-Stat projects that SSD shipments will grow at a compound annual growth rate of 154.8 percent between the years 2007 and 2013 (as opposed to just 4.8 percent for hard drives). With such disparate growth rates, it appears highly likely that SSD deployments will likely one day catch up with and overtake hard drive installations.
Selecting the Best Data Center Location
You’re no doubt familiar with the famous real estate mantra, “Location, location, location.” Well, what’s true for homes also applies to data centers. Location plays a major role in data center cost and service, so when searching for a home for your IT systems you’ll want to carefully consider where the facility is located as well as what it has to offer.
In the past, businesses generally selected a data center facility that was located within reasonable traveling distance of their main business site. But thanks to ongoing advances in data center automation and remote management, proximity isn’t as important a consideration as it used to be. In fact, many organizations intentionally use a back up facility located in another city or state simply because it provides added insurance against a local disaster bringing operations to a standstill. For those increasingly rare times when a switch needs to be physically flicked or a disc has to be swapped, look for a data center that offer free Remote Hands service. Here’s what to look for:
- Natural Environmental. Most backup systems are only useful for a few days (at best). Yet earthquakes, tornados, hurricanes, floods and other catastrophes can knock a data center out commission for weeks, or even months, as the local infrastructure collapses and is then gradually rebuilt. So when selecting a data center, think about the facility’s exposure to destructive natural forces.
- Local Environment. Pricey real estate, high taxes and other unfriendly business conditions usually don’t have a big impact on data center services, yet these factors can drive costs through the roof. Take this into consideration when evaluating data centers in different places.
- Local Utilities. Data centers rely on local utilities, particularly power companies, for essential services, so it’s important to check the quality and cost of local electric services before you commit to any particular facility. Additionally, backup power systems notwithstanding, you probably don’t want to risk deploying your equipment in an area that experiences brownouts and blackouts each summer.
- Carrier Proximity. To maximize your communications options, as well as to ensure high reliability and performance, make sure that the data center offers close and direct proximity to a wide selection of top telecom carriers.
- Skilled Employee Base. The cheapest and safest place to build a data center is probably way out in the country, many miles away from expensive cities. But a data center located in the middle of nowhere will have a tough time finding and connecting to the vital resources it needs, including the skilled people required to keep the facility running at peak efficiency. Locating your data center in an average city with a nearby major university and an educated workforce can help you keep costs down without heading to the boondocks.
- Physical Space. If a data center is already bursting at the seams with no room for future expansion, it may not have space available for you when you’re ready to add new systems.
- Construction. How a data center is constructed is as important as its placement. The best facilities are the ones that have been custom designed and tailored for data center use. Also pay close attention to the building techniques and materials used. Thick steel-reinforced concrete walls, an insulated roof to minimize heat gain and a tight, insulated building envelope will reduce cooling requirements, creating an energy efficient data center. Energy efficiency leads to lower operating costs.
You’re no doubt familiar with the famous real estate mantra, “Location, location, location.” Well, what’s true for homes also applies to data centers. Location plays a major role in data center cost and service, so when searching for a home for your IT systems you’ll want to carefully consider where the facility is located as well as what it has to offer.
In the past, businesses generally selected a data center facility that was located within reasonable traveling distance of their main business site. But thanks to ongoing advances in data center automation and remote management, proximity isn’t as important a consideration as it used to be. In fact, many organizations intentionally use a back up facility located in another city or state simply because it provides added insurance against a local disaster bringing operations to a standstill. For those increasingly rare times when a switch needs to be physically flicked or a disc has to be swapped, look for a data center that offer free Remote Hands service. Here’s what to look for:
- Natural Environmental. Most backup systems are only useful for a few days (at best). Yet earthquakes, tornados, hurricanes, floods and other catastrophes can knock a data center out commission for weeks, or even months, as the local infrastructure collapses and is then gradually rebuilt. So when selecting a data center, think about the facility’s exposure to destructive natural forces.
- Local Environment. Pricey real estate, high taxes and other unfriendly business conditions usually don’t have a big impact on data center services, yet these factors can drive costs through the roof. Take this into consideration when evaluating data centers in different places.
- Local Utilities. Data centers rely on local utilities, particularly power companies, for essential services, so it’s important to check the quality and cost of local electric services before you commit to any particular facility. Additionally, backup power systems notwithstanding, you probably don’t want to risk deploying your equipment in an area that experiences brownouts and blackouts each summer.
- Carrier Proximity. To maximize your communications options, as well as to ensure high reliability and performance, make sure that the data center offers close and direct proximity to a wide selection of top telecom carriers.
- Skilled Employee Base. The cheapest and safest place to build a data center is probably way out in the country, many miles away from expensive cities. But a data center located in the middle of nowhere will have a tough time finding and connecting to the vital resources it needs, including the skilled people required to keep the facility running at peak efficiency. Locating your data center in an average city with a nearby major university and an educated workforce can help you keep costs down without heading to the boondocks.
- Physical Space. If a data center is already bursting at the seams with no room for future expansion, it may not have space available for you when you’re ready to add new systems.
- Construction. How a data center is constructed is as important as its placement. The best facilities are the ones that have been custom designed and tailored for data center use. Also pay close attention to the building techniques and materials used. Thick steel-reinforced concrete walls, an insulated roof to minimize heat gain and a tight, insulated building envelope will reduce cooling requirements, creating an energy efficient data center. Energy efficiency leads to lower operating costs.
Disaster and the Data Center
How would you cope if your data center was severely damaged or even obliterated by a disaster? It’s unsettling to think about the many threats that could potentially wipe out your organization’s IT resources. Yet right now, as you’re reading this article, there are data centers struggling to recover from the devastation caused by fires, floods, explosions, wind, earthquakes and a variety of other natural and man-made threats. The idea that your data center could one day find itself in the same position is hardly inconceivable.
The stakes are extraordinary high. According to global insurance provider Axa, 80 percent of small- and medium-sized enterprises (SMEs) affected by a major incident either never re-open or close within 18 months. Larger firms, meanwhile, stand to suffer serious, perhaps permanent, losses in revenue, customers and reputation. Fortunately, simply by planning ahead, you can ensure that your company’s IT systems will be able to stay online and play an active role in helping your business regroup and rebuild after a calamity. Here’s how to get started:
Develop a Data Center Recovery Plan. Most enterprises already have an organization-wide disaster recovery/business continuity plan. Yet such documents often fail to adequately describe the data center’s role in supporting business operations before, during and after a crisis. That’s a major oversight, and the reason why data centers need their own disaster recovery/business continuity blueprint. The document should include information on risk analysis, budgeting, testing, deployment and other crucial issues.
Create a Backup Facility. Unless your enterprise’s data center is already located at an efficient, secure and reliable off-site facility, it’s probably vulnerable to the same perils that threaten to damage or destroy your headquarters. Most major businesses answer this challenge by creating a backup data center that’s prepared to go into action instantly on a fail-over basis. The secondary site is usually located some distance away—typically in another county or state–from the primary facility.
Consolidate Your Data Center. Many enterprises balk at the prospect of creating a backup data center, mostly due to the cost and trouble of duplicating key systems and synchronizing data. Many businesses can successfully address this problem by moving their primary data center to an off-site facility and then centralizing IT operations there. If the organization selects a place that’s virtually immune to major natural calamities (such locations really do exist within the U.S.), and that features advanced fire control, access control and other protective measures, it will have taken a major step toward ensuring its ability to survive a major crisis. The firm will also be able to take advantage of the cost and operational benefits that come with basing a data center in a facility that’s been specifically designed to enhance IT operations and to protect systems and their users from mundane interruptions, such as blackouts and network failures, as well as once-in-a-lifetime disasters.
Consider Virtualization. Savvy businesses are turning to virtualization as a way of greatly reducing the cost of restoring IT activities in an emergency. With virtualization, it’s possible to encapsulate an operating system, an application and its data into the equivalent of an application running on top of an operating system. The encapsulated application can then be easily transmitted to an off-site location—just as one would transmit a data file—and made available on a remote machine for employees to access. This capability can slash the length of downtime during a crisis from days to hours. If your organization has already adopted virtualization, it’s ahead of the game in disaster recovery and other critical IT areas. If not, you now have another reason for jumping onto the virtualization bandwagon.
How would you cope if your data center was severely damaged or even obliterated by a disaster? It’s unsettling to think about the many threats that could potentially wipe out your organization’s IT resources. Yet right now, as you’re reading this article, there are data centers struggling to recover from the devastation caused by fires, floods, explosions, wind, earthquakes and a variety of other natural and man-made threats. The idea that your data center could one day find itself in the same position is hardly inconceivable.
The stakes are extraordinary high. According to global insurance provider Axa, 80 percent of small- and medium-sized enterprises (SMEs) affected by a major incident either never re-open or close within 18 months. Larger firms, meanwhile, stand to suffer serious, perhaps permanent, losses in revenue, customers and reputation. Fortunately, simply by planning ahead, you can ensure that your company’s IT systems will be able to stay online and play an active role in helping your business regroup and rebuild after a calamity. Here’s how to get started:
Develop a Data Center Recovery Plan. Most enterprises already have an organization-wide disaster recovery/business continuity plan. Yet such documents often fail to adequately describe the data center’s role in supporting business operations before, during and after a crisis. That’s a major oversight, and the reason why data centers need their own disaster recovery/business continuity blueprint. The document should include information on risk analysis, budgeting, testing, deployment and other crucial issues.
Create a Backup Facility. Unless your enterprise’s data center is already located at an efficient, secure and reliable off-site facility, it’s probably vulnerable to the same perils that threaten to damage or destroy your headquarters. Most major businesses answer this challenge by creating a backup data center that’s prepared to go into action instantly on a fail-over basis. The secondary site is usually located some distance away—typically in another county or state–from the primary facility.
Consolidate Your Data Center. Many enterprises balk at the prospect of creating a backup data center, mostly due to the cost and trouble of duplicating key systems and synchronizing data. Many businesses can successfully address this problem by moving their primary data center to an off-site facility and then centralizing IT operations there. If the organization selects a place that’s virtually immune to major natural calamities (such locations really do exist within the U.S.), and that features advanced fire control, access control and other protective measures, it will have taken a major step toward ensuring its ability to survive a major crisis. The firm will also be able to take advantage of the cost and operational benefits that come with basing a data center in a facility that’s been specifically designed to enhance IT operations and to protect systems and their users from mundane interruptions, such as blackouts and network failures, as well as once-in-a-lifetime disasters.
Consider Virtualization. Savvy businesses are turning to virtualization as a way of greatly reducing the cost of restoring IT activities in an emergency. With virtualization, it’s possible to encapsulate an operating system, an application and its data into the equivalent of an application running on top of an operating system. The encapsulated application can then be easily transmitted to an off-site location—just as one would transmit a data file—and made available on a remote machine for employees to access. This capability can slash the length of downtime during a crisis from days to hours. If your organization has already adopted virtualization, it’s ahead of the game in disaster recovery and other critical IT areas. If not, you now have another reason for jumping onto the virtualization bandwagon.
How to Avoid Colocation Hidden Charges
Everybody loves a deal, including colocation customers. But as with any type of too-good-to-be-true business proposition, a “bargain” colocation package can actually cost far more over time thanks to hidden charges.
How can you protect yourself from falling for a colocation contract that will eventually turn into a money pit? Consider these points:
Networks and carriers. The devil is in the details. Look carefully at the number and types of networking services being offered. Remember that if a provider is committed to only a single carrier, you stand to lose a considerable amount of connectivity bargaining power when negotiating network service terms. A multiple carrier environment creates a competitive environment that drives costs down. Additionally, some carriers require a local loop fee, costing anywhere from $1,500 to $15,000. Teaming up with a colocation provider that has each carrier’s equipment installed on-site means that the local loop fee could suddenly disappear, providing a potential huge cost savings. Colocation customers in a multi-carrier environment also benefit from a blended solution of multiple fiber backbone providers. A blended environment provides a highly reliable Internet connection and significantly reduces the potential for downtime.
Cross connects. A colocation customer should never have to pay extra for interconnection. Monthly cross-connect charges are a major revenue source for most providers, so free cross connects creates a huge savings for customers—a benefit that multiplies over time. Check the fine print to see how much you will be charged per cross connect, per month. Quality colocation providers do not charge for cross connects and include a full range of commonly used cables, including single or multi-mode fiber, category 3, 5 or 6 twisted pair cable and 734/735 COAX with BNC connectors.
Remote Hands. Remote Hands service is another area where some colocation providers try to squeeze extra money out of their customers. Having to visit your colocation environment to simply flip a switch or to swap a disc is a great way of burning away time and money. So walk away from any colocation deal that doesn’t include free Remote Hands service.
Cages. If you decide to use a cage, make sure that it can be customized to meet your requirements rather than what the provider thinks you need. A cage that’s inadequately sized, wastes space or promotes inefficiency represents a hidden cost. For example, a customer may request a cage that can support 150kW of power, but receive a quote for a 400 square foot cage. While the quote may seem lower than expected, the estimate would require a power density of 375 watts per square foot (a level most providers can’t achieve). The alternative is for the provider to eventually sell the customer more space. Therefore, it’s important to ensure that the provider can honestly accommodate your real-world power needs rather than simply trying to lure you into a small cage that only appears to be a bargain.
Redundancy and uptime. Many colocation providers oversubscribe their critical infrastructure. You certainly won’t be saving any money if a vital support system suddenly fails—perhaps for an extended length of time. Make sure the colocation provider has adequate backup power generation, UPS, chiller and air handling capacity to meet your needs and the needs of the provider’s other customers. Expect a 100% Uptime SLA that has real service level objectives – don’t just take the provider’s “best effort” at keeping your IT environment up and running 24×7.
Technology expertise and availability. Look for a provider with a highly trained staff that’s available on a 24×7 schedule. Remember – a provider with years of experience will be able to operate more efficiently and cost effectively than a provider without the experience, creating customer savings.
Efficiency. Is the provider doing all it can to cut energy and cooling costs, such as using variable frequency drives, ultrasonic humidification, sealed cabinets and LED lighting? Does the provider have a peak-load shaving strategy to reduce energy consumption during the day when energy costs are typically higher? Lower energy bills translate into potential cost savings.
Geographic location. A provider’s physical location can have a major impact on what you will eventually pay. Costly real estate, high business taxes and fees and expensive utility rates are all passed along to the customer. In addition, some states may be at higher risk of natural or man-made threats, which could impact uptime. A protracted period of downtime could end up costing you more than you expect, both in terms of lost revenue and business opportunities.
Don’t zing yourself. Many colocation customers create their own hidden costs by installing inefficient, outdated, power hungry and unreliable equipment. Money invested in quality gear is rarely wasted. Consider using virtualization and other efficiency-efficient technologies to drive down costs.
Everybody loves a deal, including colocation customers. But as with any type of too-good-to-be-true business proposition, a “bargain” colocation package can actually cost far more over time thanks to hidden charges.
How can you protect yourself from falling for a colocation contract that will eventually turn into a money pit? Consider these points:
Networks and carriers. The devil is in the details. Look carefully at the number and types of networking services being offered. Remember that if a provider is committed to only a single carrier, you stand to lose a considerable amount of connectivity bargaining power when negotiating network service terms. A multiple carrier environment creates a competitive environment that drives costs down. Additionally, some carriers require a local loop fee, costing anywhere from $1,500 to $15,000. Teaming up with a colocation provider that has each carrier’s equipment installed on-site means that the local loop fee could suddenly disappear, providing a potential huge cost savings. Colocation customers in a multi-carrier environment also benefit from a blended solution of multiple fiber backbone providers. A blended environment provides a highly reliable Internet connection and significantly reduces the potential for downtime.
Cross connects. A colocation customer should never have to pay extra for interconnection. Monthly cross-connect charges are a major revenue source for most providers, so free cross connects creates a huge savings for customers—a benefit that multiplies over time. Check the fine print to see how much you will be charged per cross connect, per month. Quality colocation providers do not charge for cross connects and include a full range of commonly used cables, including single or multi-mode fiber, category 3, 5 or 6 twisted pair cable and 734/735 COAX with BNC connectors.
Remote Hands. Remote Hands service is another area where some colocation providers try to squeeze extra money out of their customers. Having to visit your colocation environment to simply flip a switch or to swap a disc is a great way of burning away time and money. So walk away from any colocation deal that doesn’t include free Remote Hands service.
Cages. If you decide to use a cage, make sure that it can be customized to meet your requirements rather than what the provider thinks you need. A cage that’s inadequately sized, wastes space or promotes inefficiency represents a hidden cost. For example, a customer may request a cage that can support 150kW of power, but receive a quote for a 400 square foot cage. While the quote may seem lower than expected, the estimate would require a power density of 375 watts per square foot (a level most providers can’t achieve). The alternative is for the provider to eventually sell the customer more space. Therefore, it’s important to ensure that the provider can honestly accommodate your real-world power needs rather than simply trying to lure you into a small cage that only appears to be a bargain.
Redundancy and uptime. Many colocation providers oversubscribe their critical infrastructure. You certainly won’t be saving any money if a vital support system suddenly fails—perhaps for an extended length of time. Make sure the colocation provider has adequate backup power generation, UPS, chiller and air handling capacity to meet your needs and the needs of the provider’s other customers. Expect a 100% Uptime SLA that has real service level objectives – don’t just take the provider’s “best effort” at keeping your IT environment up and running 24×7.
Technology expertise and availability. Look for a provider with a highly trained staff that’s available on a 24×7 schedule. Remember – a provider with years of experience will be able to operate more efficiently and cost effectively than a provider without the experience, creating customer savings.
Efficiency. Is the provider doing all it can to cut energy and cooling costs, such as using variable frequency drives, ultrasonic humidification, sealed cabinets and LED lighting? Does the provider have a peak-load shaving strategy to reduce energy consumption during the day when energy costs are typically higher? Lower energy bills translate into potential cost savings.
Geographic location. A provider’s physical location can have a major impact on what you will eventually pay. Costly real estate, high business taxes and fees and expensive utility rates are all passed along to the customer. In addition, some states may be at higher risk of natural or man-made threats, which could impact uptime. A protracted period of downtime could end up costing you more than you expect, both in terms of lost revenue and business opportunities.
Don’t zing yourself. Many colocation customers create their own hidden costs by installing inefficient, outdated, power hungry and unreliable equipment. Money invested in quality gear is rarely wasted. Consider using virtualization and other efficiency-efficient technologies to drive down costs.
