Wednesday, December 13, 2017

Mobile First–Cloud First’ Strategy – How About System Center – 07 – SCOM

Advice to the reader
This posting is part of a series of articles. In order to get a full grasp of it, I strongly advise you to start at the beginning of it.

Other postings in the same series:
01 – Kickoff
02 – SCCM
03 – SCOrch
04 – SCDPM
05 – SCSM
06 – SCVMM

In the last posting of this series I’ll write about how System Center Operations Manager  (SCOM) relates to Microsoft’s Mobile First – Cloud First strategy. Like SCOrch, SCSM & SCVMM, SCOM isn’t going to the cloud…

This product has covered many miles, first started as Microsoft Operations Manager (MOM). Mind you, Microsoft didn’t develop it themselves, instead they bought the rights of it in early 2000 from NetIQ.

Even though MOM had a refresh in 2005 (branded MOM 2005), the product had some serious issues. As such Microsoft rewrote MOM from the ground up, resulting in the release of System Center Operations Manager (SCOM) 2007 in the same year as the name implies.

From that year on SCOM got an ever growing install base, driving sales resulting in huge investments in SCOM. SCOM 2007 had some serious bugs and soon SCOM 2007 R2 was released, followed by SCOM 2012, SCOM 2012 SP1 and SCOM 2012 R2. This release cadence ran from 2007 (with the release of SCOM 2007 RTM) up to the end of 2013 (release of SCOM 2012 R2 RTM).

During this release cadence, every updated version of SCOM contained many improvements, like more speed, extended monitoring depth and breadth and better visualizations. One noteworthy feat is the integration of monitoring of non-Microsoft based workloads (Linux\Unix). The story goes that this decision was finally made by Microsoft’s former CEO, Steve Ballmer…

Azure and SCOM 2016
Up to SCOM 2012 R2 there was a big budget and good resource allocation, driving SCOM to new heights. In 2010 I went to my first MVP Summit ever, and visited Microsoft Building 44. Back then it was THE place to be, because it housed all Microsoft employees working on SCOM:

However, in the later years, Microsoft’s new future started to take more shape, moving away from software developer to the role of service provider with Azure at the center of its new focus.

Up to the release of the System Center 2012 R2 stack this new focus didn’t cause too many side effects on the on-premise product line of Microsoft.

Things started to get in overdrive however when Satya Nadella succeeded Steve Ballmer in 2014. With an impressive track record as the senior vice-president of Research and Development for the Online Services Division and vice-president of the Microsoft Business Division, he knew the power of the cloud.

Satya Nadella enfolded the Mobile First – Cloud First strategy, making clear that anything else come second (at it’s best…).

The result of this new strategy clearly shows in the release of System Center 2016, containing SCOM 2016. Whereas other new releases (SCOM 2007 > SCOM 2007 R2 > SCOM 2012 > SCOM 2012 SP1 > SCOM 2012 R2) were really upGRADES, the SCOM 2016 release is actually nothing more but an upDATE.

Window dressing?
Like removing the SCOM 2012 R2 boiler plates and replacing them with the SCOM 2016 boiler plates. Sure, some SCOM 2012 R2 components got better, but it’s more like ‘work in progress’, like the SCOM 2016 Web Console, which partially dropped the Silverlight dependence…

On top of it all, the development of SCOM has moved to India. Please don’t get me wrong, since the people in India working on the development of SCOM are very smart and bright. But they are working with fewer people compared to the ‘old days’ and do have a far lesser budget available.

And this shows. Already there is Update Rollup #4 available but still the SCOM Web Console has the so much *loved* (cough) Silverlight dependency. And SCOM 2016 is already out for more than a year…

Still going strong?
For sure, SCOM 2016 still has a lot to offer. None the less, it’s based mostly on previous investments. Perhaps the new release cadence for System Center 2016 (as such for SCOM 2016 as well), to be expected in 2018, will bring relief and a clearer vision.

This new release cadence will align more to the Windows Server semi-annual channel. Hopefully Microsoft will deliver on its promise that the first release wave will focus on SCDPM, SCVMM and SCOM.

Until then, the roadmap of SCOM is unsure, as is it with the rest of the System Center stack, SCCM excluded.

For now: OMS isn’t SCOM. OMS is all about (enhanced) log analytics, enriched with certain solutions enabling web service application monitoring. Yet, OMS is still a far cry from the enriched monitoring offered by SCOM.

For instance, alerting in OMS is quite a challenge. Also monitoring in OMS is stateless, simply because it doesn’t detect objects and doesn’t contain anything like a health model.

Sure, OMS could/should deliver monitoring in a different manner, thus making objects obsolete, but until now there are no signs of this new approach.

Therefore, based on todays world, OMS isn’t SCOM. Sure you can combine both but they can’t replace one another.

Back burner = possibilities for non-Microsoft solutions
Sure, I would love to see otherwise. But the world is moving on and Microsoft has decided to put SCOM on the back burner without offering other real monitoring alternatives by themselves. This creates a gap which other companies are more than happy to fill.

Of course, Microsoft’s marketing department tries to sell OMS as the ‘one-size-fits-all’ solution, covering everything. But reality tells us a different story all together. Combined with the ever changing pricing and licensing schemes for OMS, makes it even a harder sell.

Perhaps I am missing the bigger picture here, but this is what I see and experience from my perspective. Don’t be afraid to share your thoughts/experiences here. Feel free to comment on this posting.

Verdict for SCOM
SCOM isn’t going to the cloud at all. Sure you could install SCOM on Azure based VMs. But that isn’t the point. SCOM won’t be ported into a cloud based version. Nor is OMS at this moment capable of replacing the monitoring functionality of SCOM.

And this makes me wonder. Feels like Microsoft is turning away from a good product, without offering a real cloud based alternative. OMS doesn’t cut it yet as a monitoring solution. Perhaps later on this functionality will be added, but even then it’s important to see how it works out and what one has to pay for it.

All this doesn’t mean SCOM is dead in the water either. SCOM is still supported by Microsoft and new releases are in the pipe line. 2018 will show what the earlier mentioned release cadence is really like. Hopefully Microsoft is truly going to deliver here with TRUE upgrades instead of shameless boiler plate replacements…

Despite all of this it’s clear that SCOM isn’t going to be around for tens of years. Sure like the rest of the System Center 2016 stack it has Mainstream Support till 11th of January 2022. Until then updates, patches and the lot will come out. But after that? I have no idea.

Running SCOM 2012x? Upgrade to SCOM 2016.
When you’ve got a SCOM 2012x environment in place, changes are your company has already paid for the System Center 2016 licenses. In situations like this it always pays off to upgrade to SCOM 2016 and later.

SCOM is still a strong monitoring solution, capable of covering heterogeneous and hybrid environments, with a strong capability of customized monitoring.

Not running SCOM but looking for a monitoring solution?
However, when not running SCOM and looking for a monitoring solution, I recommend to compare SCOM 2016 with alternatives which have clearer road maps.

While you’re at it, make sure the monitoring solution offers coverage of hybrid workloads, meaning cloud and on-premise. With the shift to the hybrid world, network connections do get even more important. Therefore comprehensive network monitoring (not only limited to the device, but the flow as well) is crucial as well.

Many times companies end up with heterogeneous monitoring solutions in order to cover all their monitoring requirements. And most of the time, those solutions aren’t Microsoft based.

Recap of previous System Center stack verdicts
For more details, read the related postings.

- SCCM: Alive & kicking
- SCOrch: Dead in the water
- SCDPM: Moving into Azure
- SCSCM: Abandon ship!
- SCVMM: For now okay, but in time moving to Azure.


Ehrnst said...

Still a big advantage for SCOM is the platform it self. It has huge capabilities in the management packs, altough not very simple to create, but still powerful.
Apart from OMS/Log Analytics probably will have richer customization down the road, i havent found any product that does object based monitoring as good as scom, and still have the extension capabilites.

Wilson328 said...

Agree with most everything stated in this only has to look at the number of SCOM sessions offered at the last Ignite conference (none). It still boggles my mind that it took them almost a decade to implement something as basic as scheduled maintenance mode. SCOM admins have been clamoring for that feature for ages. I also am disappointed in what they have done with the Bluestripe acquisition. Instead of integrating that application-mapping functionality into SCOM they made it into an OMS solution instead. Application-Mapping and Transaction monitoring could have really taken SCOM to the next level.

Michiel Wouters said...

Great writeup Marnix. Agree. Although SCOM has no real replacement as an object based monitoring system, so far , it could have been much more than what it is now. Now all eyes are on OMS.

Unknown said...

Marnix, This was a great series, I thoroughly enjoyed every piece, especially this one on SCOM. It is obvious that MS isn't investing in this product, and a shame since it is so vast and useful. In my experience, most organizations are not interested in the best available products, nor even running them in the best possible manner. They balance risk vs. benefit vs. cost, and, as we in the "engine room" understand, most holders of the purse strings do not concern themselves with how well something is run, but instead IF it runs, and how much it costs. As long as the engine is running, they pay no attention. SCOM fits that category: do we really need to put such time, money, and effort into a product that just watches the shop, especially if it is all in the cloud? Doesn't MS keep an eye on that? Relating just how important such functions are can be a challenge, when the short-term bottom line is of such high priority... especially when nothing goes wrong (thanks to us watching the shop!). It is when that engine stops running without warning that all these efforts can be appreciated... or not.

Joe H. said...

So, as someone who doesn't currently have SCOM installed, is it still pertinent to install SCOM on-prem, then connect it to OMS, or should I focus on OMS only? We have the beginnings of migration talks to Azure, but nothing really in place, besides a single website with it's database.

Marnix Wolf said...

Hi Joe.

OMS and SCOM are different ‘beasts’. When running many on premise workloaads for 5 to 10 years, SCOM is still a viable solution, moreover with SCOM 1801and the semi-annual releasy cycle (SAC). Howeverm whne runing all your workloads in the cloud, OmS could be a good replacement. When running hybrid, both solutions could work for you.