Why might DC systems
be vendor-neutral?

The DC systems may, however, become vendor-neutral in the future, again because of both business and technical forces.

First, CSPs want efficient, automated operations of wide swaths of their network. CSPs cannot afford to keep their current siloed operations structures, with specialists in network and IT doing much of the work. They need to achieve their 10x aspirational goals: 10x times the agility, 10x the speed, at 1/10th the cost. This will require massive automation efforts, requiring that the domain controllers and cross-domain controllers control wide swaths of their networks without human intervention. Proprietary islands speak against that.

Second, standards and dynamic service descriptors are making the problem easier. The challenge of characterizing the capabilities of the network elements is getting easier as dynamic service descriptors and intent-based networking structures move beyond the current YANG and NETCONF models. Vendors are learning how to compile these service descriptors into their systems to instantiate features to control a multiplicity of vendors’ elements.

Main suppliers of DC systems

NEMs, SIs, ISVs and even CSPs themselves (for example, AT&T’s ECOMP and the open source version, ONAP) are all playing in the DC and Cross-domain orchestration markets. But which will predominate in the future? It is not yet clear: each has a separate set of problems in effectively providing this function.

Network element vendors face information challenges in providing multi-vendor DCs. The same problem that NEMs had in EMSs is again in play here. Getting detailed information from other vendors in a timely manner has been and continues to be difficult. But standards and technology are making that easier, as described above.

In addition, independent software vendors face margin challenges. NEMs, again, are pricing their DC systems low, bundled with their network elements (whether virtual or physical). This presents a business problem to the ISVs, especially in the DC area, although less so in the cross-domain orchestration area.

And, CSPs face the single-user cost challenge: those that decide to build their own systems have the problem of having to bear the cost of their systems and of the adaptation to the myriad of network elements. Open source initiatives are helping here, making it more cost effective.

Stay tuned

There are not strong market splits among the players, nor is it evident whether the DC systems will be primarily multivendor or single vendor. At the moment, it generally appears that the NEMs are holding the strong cards in the DC system area and the ISVs in the cross-domain orchestration system areas. But the market will decide as this area matures. ACG Research will be following these changes in its research. Stay tuned.