Exchange Server Hosting Houston

By admin  

Vulnerabilities Data Center: Why are the modern data center is missing?

The year 2008 began with a serious prognosis Subodh Bapat, vice president in the eco-computing team at Sun Microsystems, when he declared, "you will see a massive failure in a year." He continued, "We will see a data center failure of the scale", referring to the worm, which took 5% of worldwide UNIX boxes 1988 ¹.

This time it is citing security flaws as the root cause failure, but rather caused by the massive computing power required to run current applications.

While certainly an extreme position, the last year has seen a rash of data center failures that questioned why as reliable data centers are only for delivery of mission-critical applications.

The vulnerabilities ranging from the mundane, such as disasters nature and lack of infrastructure (data center outage, pipes burst, damaging the construction of fiber lines) fault hardware, storage or lack of databases and common software problems have caused periodic interruptions to business and have a high price.

Recent events in the news to support the fact that even with good planning, resource allocation and design, some of the most sophisticated may still experience a catastrophic failure.

Last summer, the state of the art-365 Data Center in San Francisco – built with more than $ 125 million – was in line for hours due to a power grid of Pacific Gas & Electric to put an important part of San Francisco in the dark. Subsequently, backup generators at the facility also experienced failure and had to be initiated manually. ²

"When the investigation of data centers, new facilities often have N +2 redundancy levels, "says Roger Smith, VP Operations Ceryx Inc." However, how are you these facilities are filled and age, which often becomes N +1, or in some areas there is redundancy at all. "

According to Sun Microsystems Executives typical life cycle of a data center is only 10 to 12 years and many data centers – built in the early dot-com era – now have to be rebuilt.

"As the person who is responsible for uptime that have to balance the applications that are deemed critical by top management and communicate clearly the cost and investment required to provide high availability, "says Roger Smith." In presenting the facts, it is clear to everyone that a data center the company could not provide the necessary levels of redundancy and even at a level of co-location would be redundancy. "

In many cases, no contingency plan could avoid the problems faced by individual data centers. On 14 July this year, Peer A data center in downtown Vancouver – One of the largest facilities in Canada – was offline for almost a whole day. An underground fire caused massive power outages throughout downtown Vancouver. Although backup generators on a pair began smoothly, the cooling system based on water as firefighters – in an attempt to extinguish the fire – exhaust pressure of water needed to maintain cooling systems operational. This caused the backup generators to overheat and any failover to UPS limited to a short battery life. ³

In a similar event this summer, the Metro, a leading hosting provider in Houston, saw a large explosion in the data center, with over 9000 customers online servers for several days. backup generators worked perfectly, but once again the fire department would not allow the installation to regain power until it was considered safe. In some cases they were physically migrated servers to a new facility.

In the aftermath of this disaster, the planet was applauded for its response to the crisis, the allocation of all resources that could make address the problem proactively and communication of status reports and the issuance of SLA credits.

Google, whose application to enterprise customers experience multiple cuts on August 6, 1911 and 15 this year, took a more reactive, the promise of building a dashboard of communication and issuance of a general credit to all customers, whether they were affected by the blackout.

The real question remains, what is the cost of data center failure and resulting downtime for organizations? Are you covered by SLA credits? Most SLA credits reflect the cost of services rendered and almost never offered for business losses.

In the Analysis of Continuity Management Conference in 2006, agility Recovery Solutions stated that 78% of businesses suffering a disaster without a contingency plan are out of business within 2 years. And 90% of companies can not resume business operations within 5 days of a disaster are out of business within 1 year.

Clearly, some deemed most critical applications and more visible than others. Large companies feel the impact immediately upon its ERP, CRM (SaleForce.com still plagued by a court of first time over two years ago due to an error with a database from Oracle Cluster?), business intelligence or E-mail systems become unavailable.

However, with the proliferation of mobile devices and everywhere access, e-mail address clearly stands out as the first mission-critical application for today. Systems such as Lotus Notes ® and Microsoft ® Exchange to maintain a living record of the existence of an enterprise, storage of all activities, and the thought processes of an organization are its employees. It's no surprise the public companies are now required to maintain an activity log e-mail for compliance purposes.

While the vast majority of businesses rely on e-mail every day to send the contracts, proposals, budgets and most of the correspondence, most email systems have not yet reached the point of reliability that provides telephone service (99.999% or 5.2 minutes of downtime per year)

According to Osterman Research, the North American businesses more experience an interruption of mail every month – and many say it could lose more than $ 100,000 as a result of a cut of a single large e-mail ¹.

Osterman also found that the average business experiences nearly seven hours of downtime for e-mail each year and that the courts can lead to many workers to a standstill, which on average are 25% less productive during downtime email.

"Forget the fact that my billing rate was shocked if I can not access my e-mail system, "says one partner at a major law firm in North America who prefers to remain anonymous. "My image of the company was greatly tarnished when I'm working on a multi-million dollar, highly sensitive front and I have to submit a set of documents using my Hotmail account because my email system is down. Someone gets fired for it. "

Michael Osterman, he continues, "Organizations are not meeting their targets for the availability messaging system," adding that the average e-mail system experiences about 70 minutes of downtime during a typical month, which translates into 99.84% uptime. It does so the question: "Is as good as this? "¹ ¹

"> Ceryx Inc., a provider of Hosted Microsoft Exchange with data center services in Canada and the United States, doesn 't think so. They were the first in the industry to offer true 100% SLA on the basis of multi-data center architecture and software design. Customers? replicates data in real time and so much in data centers – more than 500 miles away – of So even in the case of a catastrophic failure, the primary system failure with almost no impact to the end user.

"We operate on the premise that even the best data center can and will experience failure due to circumstances beyond anyone's control, "says Dr. David Penny, CIO at Ceryx." We focus our R & D activities in maintaining high application availability and rely on our replication technology to mitigate the vulnerabilities that exist in the datacenter level. And then we make capital investments and operational requirements to run every day. "

For the past 4 years Penny and her team have worked with the enterprise messaging systems such as Lotus Notes and Microsoft Exchange, the development of technology to provide high availability. Since 2004 they have been geo-replica providing Microsoft ® Exchange 2003 services to medium and large companies that see the cost benefits and performance of the solution Ceryx.

More recently Dr. Penny and her team have been working with Geographic Clustering in Server 2008 and native Microsoft Exchange 2007 CCR (Cluster Continuous Replication) technology. What this allows is the union over a wide area network. traditional groupings, which are based on the same RAID system to continue operating correctly, are susceptible to logical corruption and certain physical corruption that spread through an entire RAID array causing complete failure. Geo-clustering eliminates dependence on redundant servers in the same set of disks thereby eliminating a common point of failure.

"Even with WAN replication we need to ensure that corruption in itself is not it? T replica, "said Dr. Penny. To do this you are using log shipping with the late application instead of block level replication, thus avoiding duplication of corruption caused by defects in application. To monitor performance in the primary system closely you can stop the bad changes that are made to the secondary system.

Beyond the physical vulnerability of a single data center is protected Ceryx against a number of people of other vulnerabilities using a single data center is exposed. "In negotiating our contract, our suppliers know how easy it is for us to move the facility, "says Roger Smith." Data is replicated and not you? I need to migrate physical servers. The migration to a new facility can occur without any impact to our customers. We can not be held hostage by a bad contract or radical increases in pricing or continued poor performance. "

Ceryx also has great flexibility in routing issue and must be the backbone down or congested, Ceryx with application servers that work at both facilities, has the flexibility to route traffic through a separate facility and network congestion potential bypass might affect operators run out of a single data center.

While there are a number of solutions on the market that provide continuity through a provisional system e-mail in case of downtime, Ceryx system is different, and that does not require the user to change settings even when the email system fails over to the secondary facility. On the other hand, things like the history of email, sent items and all calendar entries remain intact.

In this sense, the Ceryx solution is no solution of continuity, but rather a high availability solution that provides layers of redundancy, from the level of software to the level of facilities.

Hosted archiving solutions – a good plan for any company facing compliance and legal – also provides a layer of security and access to e-mail records, if the primary facility to suffer complete failure. However, these solutions do not provide business continuity or availability.

On the other hand, if the primary email provider experiences of failure due to data corruption, that data is stored may be damaged as well. Large data warehouses, even in the mailbox level, leading to corruption and the current trend of vendors selling hosted Exchange email accounts allowances mass storage is the introduction of a higher probability of data corruption and subsequent failure. A good archiving strategy can be used to keep the mailbox size manageable and therefore reduce the likelihood of corruption.

So, although very valuable in today's world of mission-critical e-mail, file an organized outdoor installation should not be confused with a multi-data center strategy. Instead of file is a backup plan and will not provide good protection needs of businesses today against the inevitable vulnerabilities there is a strategy a single data center.

These vulnerabilities are typically covered in the fine print of an SLA s facility under the term "Force Mayor "? A phrase often translated as" literal Act of God or French? translation, "Superior Force" and included a clause to avoid disruptions in services due to extraordinary circumstances beyond the control of the supplier. The circumstances – as demonstrated last year – are becoming increasingly common.

Michael Osterman concludes, in his presentation on the importance of continuity of email, that the only solution to the inevitable problems that affect the delivery of mission-critical services with a geo-replication, multi-center solution data, as offered by Ceryx.


Footnotes:

¹ CNET News: http://news.cnet.com/8301-10784_3-9828570-7.html

² Data Center Knowledge: http://www.datacenterknowledge.com/archives/2007/07/24/generator-failures-caused-365-main-outage

³ Data Knowledge Center: http://www.datacenterknowledge.com/archives/2008/07/15/vancouver-power-outage-kos-plenty-of-fish/

4 Data Knowledge Center: http://www.datacenterknowledge.com/archives/2008/06/01/explosion-at-the-planet-causes-major-outage/

5 Central Network: http://www.centernetworks.com/the-planet-data-center-fire

6 CIO Webblog: http://www.cio-weblog.com/50226711/google_manning_up_for_august_outages.php

7 London Chamber of Commerce, 2006 Study

8 The importance of messaging in the enterprise: A study of continuity-mail application e Applicationcontinuity.org, 2006

9 Webblog CIO: http://www.cio-weblog.com/50226711/salesforcecom_outage_root_cause_oracle.php

10 Bnet Business Network: http://findarticles.com/p/articles/mi_m4PRN/is_2008_July_8/ai_n27893385

¹ ¹ Appcon 2007: Application Continuity Conference, "The importance of continuity Email – (webinar http://www.teneros.com/infocenter/ available)

About the Author

Executive with North American hosting companies for over 15 years.


Post a Comment

Your email is never shared. Required fields are marked *

*
*