In the business environment today, companies – particularly those that have presence in many parts of the world – operate almost 24 hours a day. This kind of operation puts a particular strain on the computing systems. It is this concern about the ability of systems to stay up that is associated with the term availability. To be clear, all technologies carry some level of risks in terms of how the systems will performance over a period of time. In addition, disruptions, whether man-made or natural, introduce further risks that have impact the availability of systems. Organizations have responded to these uncertainties by developing business continuity strategies and processes meant to put some level of predictability on system availability with respect to all types of disruptions, including planned ones such as scheduled maintenance or regular backup processes.
Setting the level of availability to a number less than 100 percent represents an acceptance by the customer that there are risks beyond the control of the provider.
In computing the availability ranges from 99.9 percent to five 9s (99.999 percent). Whilst high availability servers are rated at five 9s or higher, storage by virtue of the need to backup production data, replication, deduplication, routine maintenance such as firmware updates etc., high availability storage are rated at 99 percent uptime.
It should be remembered that in closed IT environments, availability is nearly predictable. However, the arrival of cloud computing introduced significantly higher levels of complexity and this has impacted what is acceptable level of availability. Or has it?
For more than 15 years, vendors have used availability figures to educate the market on the importance of investing in technology and processes to reduce downtime. But the business demands and technologies have changed significantly in that period.
Data&StorageAsean approached a number of vendors to find out whether IT acceptance of what is acceptable availability has changed significantly particularly among organizations that are more tolerant of disruptions.
Karthik Ramarao, Chief Technology Officer at NetApp Asia Pacific, said that as data availability is crucial for business continuity, these availability metrics do matter. This is especially so for businesses dealing with thousands of transactions per minute, such as banks or industries where real-time data analysis is instrumental to the business like in the Formula One sport. Regardless of the nature of the business, both data protection and access to data should be taken into consideration as firms have a rising need for reliability.
ARCSERVE’s vice president of Asia Pacific, James Forbes-May notes that email, e-commerce, running as a 24x7 operation, borderless transactions were nowhere near as prevalent then as they are now.
“The whole market for business has evolved significantly and that means the cost of downtime has grown exponentially especially in the past 5-7 years. When a server crashes the impact to a business today is so difficult to quantify for an SMB right up to the Enterprise. From my experience for the past 6 years in Asia Pacific and 20 years in Europe, business continuity is always the first items on the IT spend to suffer if a project overruns on cost. It’s the ‘we’ll be OK mentality’ which is up there with crossing your fingers!,” said Forbes-May.
According to Robert Chu, vice president, HP Storage APJ, “many organizations that cannot afford any downtime need bulletproof storage for their mission-critical Converged Infrastructure – with constant access to data, even in the event of a disaster. These organizations need mission-critical six 9s reliability and advanced disaster recovery solutions with superior consolidation and TCO delivering the lowest risk and highest levels of data protection,” he added.
“Data availability can be built in several cost efficient ways. For example, one can acquire a non-stop server with all components built in, or have a system integrator build a highly available cluster. However, as the former can be rather costly and the latter might be difficult to implement, it is most ideal to get the features of the first at the cost of the second. In this case, a quoted availability value would suggest that all components have been looked into and that the ideal combination suggested above might be provided,” explained Ramarao.
The majority of businesses that make business continuity a high priority are typically the ones that need to through regulatory or compliance requirements or who have suffered significantly through the impact of data loss in previous years.
“The focus has been constantly on uptime and backing up – but if you suffer a disaster – how quickly are you going to get back up and running, there are even more hidden cost impacts there it’s not just availability, it’s also ensuring you can meet an assured recovery,” Forbes-May further explains.
For some organizations, business continuity is that expensive insurance policy that the company must pay to assure business partners, regulators, customers and shareholders that come what may, the company is protected against disruptions.
Simon Ng, Regional Director, South Asia for ProphetStor Data Services shared his view that before making a purchase, having six 9s availability is what every customer seeks to possess. He states what we all know today: vendors preach what the customer likes to hear. Willing buyer + willing seller = perfect match.
“In reality, a fast, accurate & simple recovery will eventually help the customer’s mission-critical server back running in less than 30 minutes to its last good state of snapshot image. This method of fast recovery equates to buying insurance for customer’s mission critical data when all else fails or simply said that the six 9s did not materialize,” he adds.
And perhaps this should be the focus of organizations looking for the availability holy grail: a look at the business objectives of the company and how technology supports those goals. Availability then becomes a multitude of factors – hardware, software, networks, services –that need to be looked into as one connected system.