So let’s see, a few big ticket items first in the more socio/political/economic world that may affect us Silicon Valley and Networking types:
1) Repatriation of CapitalWith the new administration starting in a few weeks one of the top agenda items is lowering the corporate tax rates and finding methods to enable large companies to bring their cash back to the US. Let’s assume that, in some form, this is going to happen - what then? Well what do companies tend to do with large amounts of domestic capital, there are four usual choices:
- Dividends - who doesn’t like a little extra cash here and there. But this is a one-time deal, and not going to be a major game-changer.
- Stock Buy-Backs - generally perceived to be worse than a dividend by the shareholder, it does help the company offset any employee’s restricted stock grants by reducing the number of outstanding shares.
- Mergers/Acquisitions/Yay! - one thing Silicon Valley companies love to do is buy other companies. By the end of 2017, IF we see successful legislation enabling capital repatriation we are likely to see the beginning of a multi-year buying spree.
- Strategic Investments - some of the largest tech players also have a habit of investing into well aligned startups that could be potential acquisitions in the future, but also to see how the market develops for that particular area. Keep a close eye on this space as well if capital repatriation happens.
2) The Cloud
I tend to disagree rather strongly with Peter Levine’s assertion that the ‘Cloud is Dead’ and ‘Edge Computing is the next big thing’. It’s great for grabbing headlines, but I find it wholly unrealistic. There were 32,000+ people at the Amazon Web services show intent on using ‘the cloud’ as their primary platform for future infrastructure.
Side note: Will we see computing at the edge? Of course. IoT is an incredibly fast growing class of devices that, so far, have very few tools developed to enable IT to effectively manage their deployment and operation. It is arguably a new frontier for systems and security management which has not been nearly as well served or covered as the phone, tablet, PC or server space has.
What is amazing about the cloud is that the architectures used by the major cloud providers are so divergent from the architectures in use at the mainstream enterprises. Cloud providers have focused on automation with everything done via well structured APIs. Mainstream enterprises are traditionally managing their networks with only an antiquated command line and, if they are lucky, the occasional SNMP MIB crawl (also rather archaic as SNMP existed in the same era as MS-DOS 6.2)
For the enterprises that are large enough or their infrastructure critical enough to adopt an in-house or hybrid strategy blending public cloud and on-premises data center architectures, a new model and architectural construct is needed. In 2017 we are likely to see the increasing awareness of this and I believe the enterprises that have made a decision to continue investing in on-premises data centers will start adopting technologies that enable them to:
- Automate their infrastructure operations enabling operators to shift from configuring to programming deterministic outcomes for their deployed infrastructure
- Policy frameworks will continue to advance that provide for both on-premises and cloud-based IT operations. The most successful of these will fall into two categories:
- Frameworks that achieve breadth - supporting a myriad of workloads, locations, packaging/deployment models, et. Think bare metal, VMs, Containers, and existing network components such as firewalls, switches, load balancers, etc.
- Point products that achieve depth - There is a lot of room to improve how we secure an IT asset, especially some of the newer developer-friendly models such as containers or end-user friendly tooling such as IoT. Both of these will exist as a ‘market’ for the next several years, but awareness will continue to grow in 2017.
In 2017 we will also see:
3) Continued Developer Dominance
The main avenue for new technology to arrive into an enterprise nowadays is not through the traditional IT procurement cycle, but through the on-boarding of open source and then commercialized offerings through the developers within the business. Additionally, I have seen many instances where the InfoSec policy team said one thing and/or the infrastructure team mandated another but the developer did what was best for their speed of application delivery or what they felt was best for the line-of-business end user of that application, disregarding InfoSec and Infrastructure’s opinions.
I don’t want to get into an argument about who is right or wrong and whether or not the policy is relevant or not. But the developer usually has pocket aces in this poker match — they support the business more directly than most InfoSec or Infrastructure teams do. So as an Infrastructure guy what does this mean to me?
In 2017 I need to focus on how I can deliver the most flexible infrastructure that will support my AppDev partners regardless of what framework or architecture they want to build their applications on.
- Will they expect some workloads to never fail and others to be running for three seconds total? Absolutely!
- Will they want Kubernetes to deploy containers to four servers in separate racks that do not exist on the same subnet yet have them all use the same IP ranges? You bet!
- Will I stay up at night suffering through pages of documentation to figure this out while shotgunning Sugar Free Red Bull and occasionally slamming my head into my desk at the over-simplification of some of these architectures? Also yes!
- Are there perfect answers? No. Is there a one size fits all? Nope. Will we have reliable infrastructure platforms delivered that we can repeatedly deploy to meet any of the above (and more) deployment challenges? Assuredly yes.
4) Hacking, Cyber, Cyber, Cyber, Oh my!
2017 is going to be a royal pain for anyone trying to secure a critical workload that exists anywhere that is not completely air gapped. In the US Civil War the casualty rate was exponentially higher than other engagements - primarily because the technology (in the form of Gatling guns and field artillery) far exceeded the tactics (linear tactics were used dating from the time of Gustavus Adolphus in Sweden in the early 1600s).
Today the technology in the form of weaponized professionally developed malware, advanced phishing capabilities, remote administration kits, and so on are far more technologically capable than the majority of the tools in use by the enterprises trying to protect themselves. I had a CIO from a major bank tell me a few weeks ago that, “We are fine. W we have a XXXXXX Firewall.’
If this is the answer you have as well, start writing a really good press release and a good breach response plan — it could save your career. Look for your next big ticket breach to, again, be aided and abetted by a lack of focus on network attached devices that are not: laptops, servers, or phones.
We will continue to see an uptick in state sponsored or at least state encouraged cyber attacks and in this year I would expect to see some restructuring of the US Cyber Command — a more clean separation from the NSA has been discussed for years, but personally I would like to see it established as a discrete uniformed service. In 1945 Eisenhower created the Air Force recognizing the strategic importance of airpower as the next major battlespace. In 2017 the US needs to recognize that a major part of every future war will be fought in cyberspace and we need to be equally as prepared and focused.
One last minor thought and prediction: Alabama will continue to dominate NCAA football until we accept they need to play the Tampa Bay Buccaneers or Jacksonville Jaguars in the BCS Bowl.
To learn about Apstra and AOS 1.1, please view the full feature list, contact us or join Apstra for an upcoming webinar on January 26, 2017.