New, powerful lower-cost software is putting HA within reach of SMBs as are innovative virtualized and hosted HA cloud solutions.
Natural disasters and increasing requirements for 24/7 operations have created a need among IBM customers for high availability solutions greater than ever before.
In our effort to take the pulse of the high availability market at the start of 2012, we wanted to stimulate people's thinking but did not want to create such a rigid set of criteria that they would overlook a trend we hadn't considered. So we gave them a list of questions and suggested they consider them and write a short analysis of where they thought the market was heading. We didn't insist they answer all—or even any—of the questions we included but asked them to use the questions as a starting point to reflect on market direction. Following is the list of questions that we sent out:
- Why does tape persist as the primary backup and recovery medium?
- How has the increase in server virtualization changed high availability?
- Do you see virtual tape machines being embraced as a DR solution in the near term in the IBM midrange market?
- Has the price of high availability solutions come down sufficiently for it to be available to SMBs?
- What would you recommend to smaller companies that may not have any disaster recovery plan in place?
- What has the growth of e-commerce meant to the high availability market?
- What are the biggest challenges your company faces in promoting your HA/DR solutions?
Below are the comments we received from seven leading figures in the IBM high availability field, who have generously shared their time and thoughts so that the rest of us can better learn what lies ahead in 2012 with respect to trends in high availability.
Executive Vice President and Chief Technology Officer
The results of a year-long study on business continuity, data protection, and data sharing have just been compiled by the Information Availability Institute (IAI). Their findings, along with expert summaries appear in a free downloadable research paper entitled "State of Resilience 2011: Navigating the New Landscape: Physical, Virtual and Cloud."
This year's report highlights the ways IT organizations are now working within an environment where a variety of tools and technologies promise to transform the corporation to enable growth and innovation.
Escalating storage needs, the explosion of user-generated data, the directive to give users 24/7 data and application availability, mobile computing, and the proliferation of servers are just a few critical challenges confronting businesses this year.
IT professionals are also under pressure to implement new technologies to support demands for mobility and instantaneous access, all while reducing vulnerabilities in their protection and recovery systems. Moreover, organizations are busy supporting heterogeneous environments—physical servers, virtualized servers, and private, public, and hybrid clouds.
Reports published prior to the most recent edition also uncovered important trends. Last year's edition, for example, downloaded by thousands of IT professionals, gathered the opinions of more than 6,500 people and focused on resilience and recovery for physical, virtual, and cloud environments.
The 2010 study revealed that tape is still widely used and companies are moving toward more flexible and cost-effective recovery technologies. It also shed light on the fact that some corporations still experience permanent and unrecoverable data loss, attributable to inherent weakness in their data recovery plans.
To download the Vision Solutions report "State of Resilience 2011: Navigating the New Landscape: Physical, Virtual and Cloud," click here.
Vice President of Business Continuity Services
Why does tape persist as the primary backup and recovery medium? IT processes are often driven by what we know and what we have historically provided the business. Tape as a storage medium is relatively economical, easy to transport, and easy to store. If a tape gets exposed to water, dust, or smoke, it may still be usable. Tapes also have the advantage of being less expensive than disk drives when large amounts of data need to be stored.
How has the increase in server virtualization changed high availability? Virtualization can provide an economical platform for higher availability by hosting multiple virtual instances of critical enterprise systems rather than provisioning additional hardware. Through virtualization, IT managers can improve systems availability with solutions that mimic traditional hardware-centric HA systems. With minimal, incremental investments, virtualization as an HA platform will eliminate duplicate hardware, provide faster failover to virtual images, and be a source of fast restarts for business applications and systems.
Has the price of high availability solutions come down sufficiently for it to be available to SMBs? The price of high availability software has decreased. Also, the cost of equipment for a backup machine may be less expensive than the price paid for the production configuration. SMBs should consider the value of a managed recovery service that includes software, hardware, alternate data center location, plan creation, and staff to perform the recovery.
What would you recommend to smaller companies that may not have any disaster recovery plan in place? Ensure that your business knows what IT can deliver and what it can't. Your business will want a solution that is available all the time in 100 percent working condition. You have invested in the infrastructure to support your production solutions. Choose your high availability software, stand behind your iSeries platform, and be confident your HA solution will work—because you have tested it. Your iSeries is always open for business because your organization is always open for business, and your customers expect it.
What are the biggest challenges your company faces in promoting your HA/DR solutions? There is a growing need for IT departments to demonstrate proven value, both internally and externally. Disaster recovery should be considered an essential methodology for ensuring business continuity and survival. Often, these goals are minimized in tough economic times. Saving pennies and risking survival dollars in a disaster does not make financial sense. Broad market visibility about managed recovery services from providers like Velocity can be a challenge when an organization's comprehensive disaster recovery plan is now more critical than ever.
Power Systems HA Product Manager
Companies embracing "smarter" computing with IBM i running on Power Systems are focused today on building IT infrastructures that enable them to deliver services to their business with both higher quality and superior economics. A key driver of improved economics with IBM i has not only been the increased use of server virtualization with PowerVM, but also deployment of storage and tape virtualization. Combined with the relentless pressure to improve service-quality levels, IBM sees a major market shift in the deployment of high availability solutions.
The trend for large enterprise users of IBM i is very clearly toward using PowerHA SystemMirror clustering in combination with storage area networks like IBM DS8000 and IBM V7000. PowerHA enables rapid and simple switching between IBM i systems in a cluster. Companies deploying PowerHA are also exploiting new storage software, including FlashCopy, which enables a near instantaneous copy of IBM i data, and Metro Mirror or Global Mirror, [which] provides a remote data copy. IBM i clients using PowerHA consider it easier to maintain, easier to role swap, and more cost-effective than logical replication solutions.
For midsized companies, simplicity and cost-effectiveness remain a top priority when considering improving business resilience. We have seen increased use of virtual tape backup both to their system storage and to cloud providers, including IBM Business Continuity and Resiliency Services. We also see PowerHA being deployed in midsized companies with internal storage, both to reduce cost and lower risk through simpler role-swap testing.
Finally, in all IBM i markets we have seen an increased demand for resiliency skills, both in our IBM Lab Services practice, which specializes in PowerHA and SAN implementations, and in our IBM Business Continuity and Resiliency Services practice, which offers a full range of services to help clients manage business risk.
High availability is becoming a standard requirement across a much broader range of companies. This means that those companies who have previously felt high availability is beyond their budget are now looking for a solution which will provide high availability without the perceived high cost. One area we are seeing growth in is the smaller companies, the ones who have been sitting on old iSeries for a long time and have finally seen the opportunity to move forward with the new Power7 pricing model. This has provided another benefit because the old system which was enough to run their business is now worthless as far as re-sale is concerned but can be used as a valuable target system for their new Power7 hardware. We have priced our high availability offering to fit into that niche and see a number of clients moving in that direction. Another valuable option is hosted high availability; the cost is spread over a period of time, and the expertise to run the high availability solution is provided by the hosting partner. We are currently looking at virtualization as another option where the new Power7 system is carved up to provide a recovery partition with minimal cost. High availability is now affordable even for those smaller users who previously felt it was beyond their reach; it just takes a bit more imagination to come up with the options.
Most companies could benefit from a high availability solution, but the age-old test is still valid. You must ask yourself, how long can I afford to be down? If it's one or two days, then probably a tape solution is all you need—providing you design it carefully and keep your backups offsite and have a method for retrieving them. But if you need a recovery that is quicker, then you will need a high availability solution.
Tape persists because it is easy and convenient for what it does well. It's not a disaster recovery solution in any immediate sense because it takes a long time to restore data. And tape is relatively inexpensive. We must keep in mind that backup solutions are different from high availability solutions. People still want to have a backup of their data that they can use for various reasons even outside high availability, so tape continues to have its place.
It's clear that e-commerce goes hand in hand with 24/7 operations. When that happens, it's not only difficult should you go down, but it is also quite costly. You want to do whatever you can to maintain operations. Needing to be up 24/7 makes high availability just that much more important. While most companies could benefit from HA, I think some people have hesitated to look at high availability because, historically, there has been a perception it is expensive, complicated, and hard to implement. That perception is changing as new solutions are introduced.
Needing a second machine to support high availability is a detractor to some, but over time, people tend to upgrade their machines. If the machine is on lease, the customer has the option to buy the machine at a reasonable price, and if they have purchased it but are now upgrading, there is always the question of what to do with the old machine. The answer is that it is still very useful as a DR machine. The systems don't have to have the same chipset or same release of the OS; they can be two releases apart. IBM also has a nice program that gives good discounts on hardware and software for a second machine being used for high availability.
So far, there hasn't been much of a price reduction in HA solutions in the marketplace except for ours—which is considerably less expensive due to our business model. It takes people awhile to catch on to the lower-cost yet equally powerful solutions available.
Convenience features such as automatic monitoring and messaging make high availability solutions easier to manage. We expect those kinds of features will help make it more attractive for companies that may not have considered an HA solution in the past.
I think SaaS-based data centers offer a viable option for some people. In certain cases, it's a good solution for customers when it pencils out right in terms of budget, since the user doesn't have to have a second machine. The data center solution also offers the advantage of storing the data off-site, which you want to do regardless. In some cases, companies running on-premise HA solutions will replicate data to different divisions. Different companies who normally cooperate have been known to offer each other off-site storage as well. Yet the data centers have to be running some kind of replication software, and some will run ours, while others run someone else's.
As far as trends, I think people are fed up with complicated solutions that don't perform as expected. I believe we're going to see an increase in convenience, automation, reliability, and economy in HA solutions. Also, IBM has a hardware solution that, despite being quite elegant, could be simplified. If they ever made that as simple to use as RAID or single-system disk mirroring, that would be slick. It generally requires clustering, and, since it's disk to disk, you can't easily select what files or libraries to replicate, but it's definitely something to watch.
Senior Vice President
Over the past few years, IBM i shops have been taking up real-time data replication for their high availability and disaster recovery solutions in increasing numbers. In 2012, we expect to see this trend continue and probably accelerate. The popularity of real-time data replication between IBM i systems has been driven by lower costs of hardware, bandwidth, and the HA software itself. It is also because having the capability to fail-over or role-swap to a backup server without an extended outage and without suffering any data loss is becoming the expected norm—not just a "nice to have."
We expect the cloud-based disaster recovery and high availability model will be a game changer, and we can see this trend developing quickly. To date, there have been few providers with some notable exceptions, such as Sungard Public Sector, offering a real and viable Cloud DR Service. At Maxava, we have been advocating cloud (or application service provider or SaaS) for many years and already have more than 100 customers relying on Maxava HA for cloud-based HA/DR. Recently, we have noticed a real change with many more cloud-based HA/DR providers offering services and many customers moving to utilize them. We expect this trend to increase, and it may finally be the catalyst for many IBM i shops to go beyond tape and secure their data in real time.
We also expect to see more development in connectivity via handheld devices, such as our maxView interfaces. 2011 was the most expensive year in history for natural disasters, and events such as the Japanese and New Zealand earthquakes highlighted the fact that people and systems may survive, but physical access to data centers may not be possible. In these circumstances, the ability to remotely switch systems can be critical. This kind of technology is also important to efficiently manage multi-tenanted hosted systems. Expect more innovation in this area—beginning with Maxava's new and improved maxView lite application for the iPad and iPhone, which can be downloaded free from the Apple App Store.
While some may think all HA/DR products are fundamentally the same, there are real differences in the underlying architectures of the various solutions. These differences significantly impact such things as performance, ease of use, and whether processing occurs on the production system or on the backup. They therefore directly impact how much data is securely "off-production" and also the speed of getting users back online when failure occurs. Whether customers decide to purchase their own HA/DR solution or use a hosted service, there are real choices in the HA/DR market in 2012, and customers will need to continue to carefully consider their options. 2012 may well be the year that "last night's tape" is no longer considered an adequate DR plan…time will tell. (Download a free copy of Maxava's HA/DR Jargon Buster here.)
IBM i HA Expert and Owner
HA will increasingly be implemented by IBM i users. Prices will continue to drop, and functionality will increase; large HA vendors are losing market share to more-agile small ISVs, and backup virtualization is gaining interest.
HA will grow because the price is dropping and the need is more immediate. More and more data is real-time, and no company can afford to lose even a couple of hours due to downtime.
2012 will be the year of small HA software vendors. Solutions like iSB/RSF-HA offer a lower price point and higher functionality. To survive, we have to offer more than the big guys. We're more nimble and can make changes more quickly. What we lack in installed base and marketing dollars we more than make up for in features and functionality.
There will always be a need for long-term data retention, and tape remains the most cost-effective medium for that. New disk-based virtual backup solutions automate and speed up backups, first to disk and then to tape or disk. As a result, companies can reduce backup windows, consolidate backups from multiple heterogeneous servers, add newer tape devices while retaining legacy data, and reduce total backup storage costs.
Finally, let's not forget two factors that won't change as long as we have humans and weather. Humans make mistakes. This January, an outage that IBM attributes to an error in UPS program logic took down a large Montreal IBM server farm, locking customers out of their bank accounts for over 24 hours and delaying Air Canada flights. Mother Nature will continue to be a force to be reckoned with, as we saw with last year's unusual U.S. tornados, earthquakes, and flooding. In 2011, five of the world's 10 largest insured loss events were in the United States, totaling insured losses of $23 billion and economic losses of $35 billion, according to the Aon Benfield insurance company.