We live in an information age, and from young to old we all want to have information at our fingertips. This does not just apply to the consumption of news or social media, but it is simply a fact that many businesses will close their doors if they lose (or lose access to) their data. However, disaster can strike at any time, and as we have seen over the last few years, natural disasters have wreaked havoc on cities and countries in ways previously unimagined.
Then along came the “cloud”. And you may be a cloud “Grinch” and not like it, but you might be surprised to find that with so many applications around us, you may already use the Cloud without even knowing it.
But this article is not about “Cloud” per se, but more about making sure you look at options to protect yourself from losing data by looking at Disaster Recovery (DR) and “Cloud”. In technology, we love the use of abbreviations. Sometimes you need to look up what some of these mean as there are so many nowadays, but Disaster Recovery as a Service or “DRaaS”, is one that I will focus a little more on here. What is this? What does it mean? Should you look at this?
Understanding the basics is important!
Even though Natural Disasters are more common and definitely something to be prepared for, there are still two others that some forget about – Infrastructure Failure and Human Error. Yes, these, especially human error, are still one of the biggest causes of disaster. But we should not focus on the type of disaster. There are five things you need to take into account when you start looking at your Disaster Recovery plan. And these are:
• Recovery Point Objective (RPO): Can be seen as the maximum amount of data loss expressed as a timeframe that is considered acceptable, which in short means, you cannot lose data older than this specified timeframe.
• Recovery Time Objective (RTO): What is the maximum time you can spend on “recovery” following a disaster, before you have to have your system operational again?
• Zero Data Loss: If disaster strikes and you lose your primary system, can you recover ALL data up to the point in time when the disaster struck? Think about this… no transaction can be lost
• Change Control: Imagine you have the disaster; where do you start? Who makes decisions to fail over to the DR site? Who should perform the failover? What happens to the applications, business processes and all the other areas? Not to mention the process to follow when doing code releases.
• Testing and Documentation: Test and test as much as you can, and document everything! Do this on a regular basis! Too many times companies implement a DR plan, yet they never test it regularly. This is as scary as not having a DR plan in the first place. What if things change and your DR plan is not tested or updated?
But it does not stop there, taking the above, we start to get the idea that there is more to Disaster Recovery. It is all these processes, policies and procedures we put in place to ensure recovery and continuity of infrastructure, applications and databases – all of which is critical to Business Continuity (BC).
You might go – wow, how can you bring in cost so quickly? Well, there is something that should be understood from the beginning, depending on some of the answers to the above questions such as Zero Data Loss, there will be a cost associated with it. In general, most environments that require strict RPO and RTO or even Zero Data Loss, will tend to have much higher costs associated with providing these solutions, than those available for less restrictive RPO and RTO requirements. That is why it is extremely important at the early stages of looking at solutions to make sure you actually talk to the business and listen closely to their requirements, do not just make assumptions.
At this point I want to use an example of something most of us are so familiar with …
Do you take the blue or the red pill? And in this example, I want to state the question:
Do you go with traditional Disaster Recovery? Or do you evaluate the Cloud and look at what Disaster Recovery as a Service can provide?
“Traditional” Disaster Recovery
When we look at how many DR solutions are being implemented, most will be familiar with the on-premise model where you either have your DR site close, if not in the same datacenter, hopefully at least in another datacenter in a different geographical location. You have to maintain both sites. When implementing this traditional solution, it tends to be something similar to the diagram below.
You go through these key phases, all of which are crucial to the successful implementation of a DR solution, but there are two parts that should be highlighted, the ordering of hardware followed by the building of the systems once the hardware arrives. And for those of us who have worked in this type of environment will know, this can take time and the upfront cost can be substantial.
DR in the Cloud (DRaaS)
Now when we start looking at Cloud, things start to get interesting. There are many cloud providers today and more and more are entering this competitive scene. They range from a small specialized service catalogue to a huge extensive catalogue that caters for almost any industry. Now as with most things, it is not a “one glove fits all”. But before I get into that, let’s first take the previous diagram and see what it can look like when using a Cloud based DR solution. This can be either a hybrid solution, where the primary site is still on premise but the DR site is Cloud based, or, if you have already invested in the cloud, you can do Cloud to Cloud DR. Note that it could even be using two different Cloud providers - also referred to as Inter Cloud DR.
From the above diagram, we can see that there are 3 key differences. Firstly, there is no huge upfront cost for ordering new hardware. Secondly, there is no building the infrastructure like setting up the systems, storage, network etc. And lastly, there is no time delay. In this scenario, you still have the traditional method of detailed planning, design, and approval phases, but the time to provision the environment in a cloud based solution is much faster and has little upfront cost. This might not seem like it, but this is actually a game changer! Companies do not have to wait to establish DR solutions and they do not have to spend huge upfront costs.
But let’s step back for a moment – is cloud really that new? Well in some cases yes, it is, but in others, is it really? There are many companies that have outsourced their infrastructure and IT environment to 3rd party companies… is this not some sort of Cloud? Yes, it is, but over the last 15 years this has evolved at a rapid pace and the playground is looking a little bit different now.
Cloud – explained in one paragraph
A tough task for sure, but think of it this way. Someone else is looking after either your Infrastructure or Software, or both! It kind of explains the “..aaS (as a Service)” that we see all the time. There are many of these and the list is growing. The core ones you will see on a regular basis and which you might already use are: Software as a Service (SaaS), Infrastructure as a Service (IaaS) and Platform as a Service (PaaS). Each of these are huge topics, but one key thing is common – someone else is looking after things for you – whether hardware or software. And most of these providers tend to have two key pricing models – Bring Your Own (BYO) license and Pay-as-You-Go (PYG) model. But it has to be said that the most popular, and growing rapidly, is the Pay-as-You-Go model. Security is always at the forefront. Everyone might say it is not secure, but in fact, most cloud providers spend huge amounts of time and money to ensure they have good security measures in place and obtain the required certifications. And to add one final sentence to this paragraph – considering these solutions allows you to focus on your business, while they manage the infrastructure.
So what about DRaaS
Why not get someone that can help you establish a DR site in the cloud, where they look after the physical and in most cases virtual servers. They can help you establish a DR site in the cloud for your database, application and even back office services. Then working closely with you, establish a failover (Switching) procedure to the cloud-based site in case of disaster – all of this within clearly defined and agreed Service Level Agreements (SLAs). Working with these providers you can establish a DR site that is tested, documented and that you can trust. Using this type of solution is ideal for Hybrid Models where companies still have heavily invested on-premise solutions.
But, and there is always a but
Do your homework. Keep in mind one key factor when looking at hybrid solutions – Networking. Even though it is getting better and better every day, there is still a lot of areas that do not have excellent network connectivity to Cloud providers. These can in many cases be upgraded but do make sure that things like bandwidth and latency are considered.
And always review the Cloud provider security options, you might be pleasantly surprised by all the hard work they have already done in this space. But do make sure that they tick all the boxes you require. Remember that shared resources are used in many cases due to Virtualization – so again, make sure you review the service catalog and SLA agreements.
One final point of interest is integration. Not all providers have extensive API interfaces to allow for integrating existing or new applications. Having these options might seem like nothing but could be a major advantage down the line if they have them available.
Costs vary from provider to provider but there is one big advantage with cloud based solutions – you pay for what you use. You reduce upfront cost, so you can start small and grow, adding as you need, and in most cases easily reducing when needed as well. This is where it is important to review service catalogs and yes… they can be complex. Things go into hourly pricing and there is lots of fine print, but be aware that long-term contracts can save a lot of money. So, compare service catalogs and costs – it is worth it.
On-premise will still be around for a long time, but cloud popularity is growing and slowly catching up. Maybe a lot faster than most of us think! A good way to test the waters is to review and dust off that old DR plan – it might be that the Cloud can help you establish something you can trust in a time of need.
In closing I would like to highlight that Cloud-based DR solutions can provide you with flexible, scalable and affordable solutions with self-service options in most cases. Allowing companies to now focus on their business, not infrastructure.
Anton Els is based in Auckland, New Zealand and is the VP Engineering at Dbvisit Software Limited. He has been working with Oracle database technology since version 7. He became an Oracle ACE in 2016 and is an Oracle Database 11g Certified Master. Anton is also a co-author of "Oracle Database 12c Release 2 Multitenant (Oracle Press)”. Anton is a regular speaker at various Oracle technology related conferences including Oracle Open World and Collaborate. Anton loves sport, but is a techie at heart and enjoys playing with technology in his spare time.