Putting the silver lining back into the public cloud

Pratyush Khare

Cloud computing has made some pretty big promises since it exploded onto the IT landscape – especially when it comes to efficiency and cost-effectiveness. And enterprises are clearly buying in. According to Gartner, public cloud spending in APAC will hit US$7.5 billion this year (2015)

Unlike the situation just a year or two ago, CIOs in almost all large enterprises – including the banks and large insurance companies – are under pressure from CFOs to upload data to the public cloud to reap the cost-saving benefits it is widely supposed to deliver.

The App challenge

One problem is the proliferation of applications within organizations. Applications, which are typically designed to talk to the storage in the data center using CIFS and NFS protocols – are not good at talking to data stored with public cloud providers, which all communicate using HTTP.

The obvious solution is to re-write the application, but that’s exactly the last thing any sane CIO wants to do. What happens if there’s a glitch and the app fails? In an important app, the downtime could literally paralyze the organization.

Some of the newest and most innovative content management solutions can resolve these issues by acting as an interpreter for the applications. This facilitates full, policy-based data tiering to a public cloud platform – whether that is Microsoft Azure, Google, or Amazon– without changing any application code. Applications essentially become cloud-ready without any need to re-engineer them.

Counting the cost

While the app issue can be dealt with. there are other considerations that in some cases make the public cloud approach effective in solving an immediate problem, but create new difficulties in future.

What CIOs know, but CFOs and other LOB executives often don’t see is that, while public cloud storage might seem cheap, accessing the stored data when it is needed can become a very expensive undertaking indeed. Especially if you are pulling down a lot of files to, for example, conduct an analytics exercise.

Recently I talked to a CIO at a major US-based retailer which does about half of its business in the holiday season (November/December). It uses certain apps for heavy analytics every January to produce sales reports, creating data storage headaches and a burden for compute resources.

The retailer decided to move this app to the public cloud and until December, the average bill fluctuated comfortably between US$90-95 thousand. Then, in January, they received an invoice for US$1.3 million! The cause of the huge bill was the LOB users – who didn’t realize that the data was now on a public cloud – and were running massive numbers of queries and reading huge volumes of data, driving up the bill.

The good news is that this is not the case with the newest content solutions because they retain the meta data – the file name, author and version control information, and even information on which customer to which the file applies – on premise. This enables it to be accessed and analyzed quickly and inexpensively.

Take an audit for example, which might require thousands of files from years back to be examined. With innovative content solutions, all of the queries can be conducted locally. That’s not only faster, but few, if any of the queries ever touch the public cloud, so the access bill is kept to a minimum.

The compression conundrum

Even when access charges are kept under control, providers still charge by the petabyte, which can get expensive quickly in these days of big data. Storage fees can feel doubly painful when many cloud providers compress the data they receive, but still bill the customer for the original data volume.

That shouldn’t be an issue for advanced content platforms that enable the data to be compressed on the premises before it is uploaded to the public cloud. In that case the users pay only for the space their data takes up on the provider’s server, not twice as much!

Ensuring security

Security has always been close to the top of the list of cloud concerns. And, in today’s post-Snowden universe, it still is.

While every reputable public cloud provider offers encryption to protect customer data, rendering it useless if it were ever accessed without authorization. However, the question many are asking is how secure are the encryption keys? Have they been compromised by hackers, or even made available to shadowy national security organizations.

It’s a thorny issue, but there is a simple solution. Encrypt the data before it leaves the data center.

The newest and most innovative content solutions offer organizations an opportunity to do precisely that, so the data is secure long before it reaches the cloud provider’s servers. It can even be encrypted again on arrival, protecting it from hackers and other miscreants, and anyone else who might be interested in doing a bit of snooping.

Freedom of choice

Moving to the public cloud is a big decision. But what happens if you want to switch providers for any reason?

The answer is that a lack of clear standards for data formats and APIs to ensure interoperability between infrastructures can make changing providers problematic. That can leave many organizations feeling locked-in.

In fact, this may be the single most important reason for implementing an up-to-date content platform based on open standards. With built-in multiprotocol access, the newest offerings eliminate the problems of API integration from the outset, so the vendor lock-in becomes a non-issue.

Of course, the cloud journey can be a long one and enterprises are continually faced with new challenges. However, the latest and most innovative content platforms offer CIOs an opportunity to put a silver lining back into the way they view and pay for public clouds.

The author is the director, Solution Sales, Asia Pacific, Hitachi Data Systems