Ten years ago, Nicholas Carr wrote a paper entitled “IT Doesn’t Matter” published in the Harvard Business Review. He might not have realized the far-reaching effects but in many IT shops, and with many senior executives, it signaled a shift from focusing on compute, storage, data centers and networks to applications. This also coincided with the rise of enterprise applications and, as a result, CIOs spend a lot of time discussing packaged applications, integration, and implementations, resulting in the treatment of the fundamental engine of their business as a commodity. But in most companies, packaged applications represent less than 20% of the overall footprint.
Unfortunately, paying all the attention to the ERP or CRM implementation has left many with a horse-and-buggy compute and storage infrastructure, when the competition is moving to cars. And the horse and buggy are getting more expensive every year. The cost of computing is not the cost of the server or the storage. It’s the cost of all of the underlying network hardware to connect the computers; plus the diesel generators, power distribution and building access hardware; plus all the software to manage power, cooling building access, firewalls, server, storage and virtualization; plus the people and software to manage the security, availability, performance and changes in the compute and storage. As a good rule of thumb, the true cost of computing is at least 4-times the purchase price of the hardware per year. And with software becoming more complex, and hardware continuing to decrease in cost, the multiple is only increasing. There can easily be an order of magnitude (10x) of difference between the best and worst compute and storage services today— hardly a commodity.
Furthermore, consider all of the software products that compose your compute, storage, data center and network. It’s probably at least 25 products. On average, every software product makes 50 changes per year, so you’re conservatively managing over 1,000 changes per year. And if you’re not staying current, then why are you paying the support bill every year?
Finally, everyone is concerned about security. You should learn a lesson from the Iranian nuclear program, which had big guard dogs protecting the centrifuges. But as you might have read in the news, a clever virus destroyed those centrifuges. So if you’re not making sure all of the software you have managing the compute, storage, data center and network is virus free, then it doesn’t matter how big your guard dogs are. But it’s not just making sure there is no bad software; you need to make sure you use the security patches issued from the vendors. We did a study of just 5 commonly used software products, and in one year, there were 137 security patches.
So if you’re not actively managing your foundation every day, your horse and buggy are getting more obsolete and more insecure.
Nicholas Carr’s article might have gotten all of us to focus on applications and take our eye off of the foundation of any modern company – your compute, storage, data center and network. Focusing on the apps has made some difference, but given there are so many custom, industry-specific applications, and the rise of new applications based on technologies born in the Consumer Internet (Hadoop, Cassandra, Solr, Hbase, etc.), there needs to be a return to paying attention to the fundamental engine. Your current horse and buggy will only get you so far. Your competition is starting to leverage the ability to dramatically lower the true cost of computing, provide an agile and current environment for all applications, and deliver it securely. Shouldn’t you be doing the same?