|Read the Digest in
You need the free
In this issue:
Browse through our Useful Links.
See our article archive for complete articles.
Sign up for your free subscription.
Visit our Continuous Availability Forum.
Check out our seminars.
Check out our writing services.
Never Agains Never End
It seems that no matter how many incidents make the press about datacenter outages, including our own Never Again articles, the stories keep repeating themselves. Upgrades go awry, failover procedures fail, power systems break down, networks become notworks, and the ever-present human error bites us. Analysts estimate that 70% of all outages are caused by or contributed to by people. Too often, a person becomes a single point of failure in a complex outage. One wrong move, and disaster happens,.
True, many IT staffs are overloaded and take shortcuts that can be dangerous. Is there a tested fallback plan in place should an upgrade cause problems? Are intranets redundant and monitored so that failures can be quickly located and isolated? Are human single points of failure backed up by another person to ensure that all actions are appropriate? No matter how many outages we read about, we still fall into the same old traps.
We at the Availability Digest are always looking for new Never Again stories that teach a lesson. If you are aware of any that we have not covered, please let us know. We depend upon this information to bring real-life issues to our seminars on high availability and to our technical writing services.
Bill Highleyman, Managing Editor
Our last several issues of the Availability Digest included many notable failures and cyber attacks that took down corporate systems. Spamhaus, a spam-site blacklisting firm, was taken down for days by a massive DDoS attack launched by one of their blacklisted sites. Two Middle Eastern banks were robbed of USD $45 million by hackers who compromised their gift cards. A phony AP tweet claiming the President had been wounded in a White House attack crashed the stock market. New York City’s new 911 system crashed four times in the first two days of operations.
During this time, several other outages worth covering occurred. These outages were caused by a wide range of problems. Interestingly, over half of them were environmental. Some were power and cooling problems, but others were unimaginable – a rat shorting out a circuit-breaker panel and a train severing a fiber cable.
Such incidents emphasize the need for a thorough, well-documented, and well-practiced business continuity plan that is independent of the causes of outages. You may think that you have all of your bases covered, but you never know what oddballs fate has in store for you.
Bring Your Own Device (BYOD) is the new gateway into corporate networks. An increasing number of employees are using their smart phones, tablets, and notebook computers to conduct their work at home or on the road by connecting outside the corporate firewall into their companies’ servers and databases. This is a great convenience for workers and represents an increase in productivity for their employers – the distinction of a fixed workday is disappearing.
Unfortunately, BYOD also means that malicious actors can gain access to a company’s network by infecting these devices, which are woefully unprotected. Mobile malware is rapidly becoming a greater concern than direct infections of the systems themselves.
Attacks against corporate networks invariably begin by stealing an employee’s credentials. When employees access the network from a device that is beyond the control of IT, the risk represents the weakest point in the network. While no single solution can protect against the constant explosion of cyber threats, IT must better protect the mobile devices used by the company’s mobile workers. Help is available from companies that specialize in mobile-device protection.
Despite previous concerns of many in the HP community that HP’s robust OpenVMS operating system may be headed for extinction, HP recently released an OpenVMS roadmap that indicates that OpenVMS support will continue for years to come. According to the roadmap, the latest version of OpenVMS, V8.4, running on Integrity i2 servers, will be provided full HP support through 2020 and perhaps beyond. The roadmap also provides continuing support for legacy versions of OpenVMS running on VAX, Alpha, and Integrity servers.
Since the late 1970s, OpenVMS clusters and active/active NonStop systems have been the gold standard for applications that require continuous availability.
HP’s OpenVMS roadmap has set date-certain schedules for ongoing support through 2020 of OpenVMS V8.4, with possible extensions if warranted. There is plenty of time to plan a strategy for continuing to run your OpenVMS systems and applications, whether it be on supported HP platforms; porting to other HP operating systems such as HP NonStop, HP-UX, Linux, or Windows; or emulating OpenVMS on commodity hardware.
Servers in a data center must not only be interconnected, but they also must connect to local users and to the Internet. Today’s internal networks use Internet technology and are called Intranets. A company’s Intranet must not represent a single point of failure if its systems are to provide high availability.
To offer high capacity, performance, and availability, the customer-facing systems are often pools of servers. Load balancers are used to distribute incoming traffic to these servers. However, a load balancer can represent a single point of failure in this critical portion of a company’s Intranet.
Load balancers from Loadbalancer.org (www.loadbalancer.org) solve this problem. Their load balancers are configured in active/passive pairs with instant failover so that customer-facing services are not interrupted should the active load balancer fail.
The Loadbalancer.org load balancer configurations provide a wide range of load-balancing options to suit most networks. They are certified for many applications and support several load-balancing algorithms. They are offered as appliances with a wide range of performance capabilities from 1.5 gbps to 10 gbps throughput and up to 7 million concurrent connections. They are also offered as virtual load balancers running under VMware or Hyper-V.
Sign up for your free subscription at http://www.availabilitydigest.com/signups.htm
Would You Like to Sign Up for the Free Digest by Fax?
Simply print out the following form, fill it in, and fax it to:
+1 908 459 5543
The Availability Digest is published monthly. It may be distributed freely. Please pass it on to an associate.
Managing Editor - Dr. Bill Highleyman firstname.lastname@example.org.
© 2013 Sombers Associates, Inc., and W. H. Highleyman